Paper deep dive
Digital Speech Acts Retain Control of Copyright with People, Not Platforms
James Golike, Ehud Shapiro
Intelligence
Status: succeeded | Model: Gemma-4-26B-A4B | Prompt: intel-v1 | Confidence: 96%
Last extracted: 6/20/2026, 4:56:52 AM
Summary
The paper argues that 'digital speech acts'âvolitional, cryptographically signed content created on personal devicesâshould be recognized as copyrightable expression under existing U.S. legal precedents like Burrow-Giles and Feist. The authors contrast centralized digital platforms, which use copyright and contract to extract user content and separate ownership from possession, with 'grassroots platforms.' These grassroots architectures use cryptographic signatures to ensure that ownership and possession coalesce in the person, preserving attribution, accountability, and authorship through a non-unbundlable provenance chain. This architectural approach is presented as a prerequisite for digital sovereignty and democratic self-governance.
Entities (8)
Relation Signals (5)
Burrow-Giles v. Sarony â establishes â Authorship
confidence 100% · Burrow-Giles locates authorship in volitional creative choices
Digital Speech Act â establishes â Attribution, Accountability, and Authorship
confidence 100% · through which the person simultaneously establishes attribution, accountability, and authorship over the signed content
Centralized Digital Platforms â extracts â User Content
confidence 100% · platforms use copyright defensively over their code and contract offensively over user content
Digital Speech Act â qualifiesas â Copyrightable Expression
confidence 95% · digital speech acts qualify for copyright protection under existing U.S. precedent
Grassroots Platforms â preserves â Copyright
confidence 90% · the digital social contract underlying grassroots platforms preserves this copyright by design
Cypher Suggestions (0)
No Cypher suggestions yet.
Abstract
Abstract:Legal precedents protect computer code as copyrightable expression. They have enabled centralized digital platforms -- operating from corporate servers that hold all user data -- to construct private governance regimes through the interaction of copyright, contract, and technical architecture: people who create virtually all platform value must surrender effective copyright control through Terms of Service agreements as a condition of participation. In contrast, grassroots platforms consist of cryptographically-identified people operating their networked smartphones independently of any server or global resource; each person holds their own data on their own device, with no third party in possession or intermediation. Here, we define the notion of a \textit{digital speech act} -- a deliberate volitional act by a person of cryptographically signing personal content with the person's private key, carried out on the person's own device -- through which the person simultaneously establishes attribution, accountability, and authorship over the signed content. We contend that (\ia) digital speech acts qualify for copyright protection under existing U.S.\ precedent: \textit{Burrow-Giles} locates authorship in volitional creative choices despite mechanical or algorithmic processes, \textit{Feist} supplies the minimal-creativity threshold, and persistent device storage satisfies the Copyright Act's fixation requirement; (\ib) the digital social contract underlying grassroots platforms preserves this copyright by design -- signed content cannot be unbundled from its signature, and the full provenance chain accumulates as content is forwarded -- so that ownership and possession coalesce in the person; and (\ic) copyright in digital speech acts is a prerequisite for digital sovereignty and democratic self-governance.
Tags
Links
- Source: https://arxiv.org/abs/2606.19263v1
- Canonical: https://arxiv.org/abs/2606.19263v1
Trouble viewing inline? Open PDF directly â
Full Text
254,419 characters extracted from source content.
Expand or collapse full text
Digital Speech Acts Retain Control of Copyright with People, Not Platforms By James Golike and Ehud Shapiro * Abstract Legal precedents protect computer code as copyrightable expression. They have enabled cen- tralized digital platformsâoperating from corporate servers that hold all user dataâto construct private governance regimes through the interaction of copyright, contract, and technical architec- ture: people who create virtually all platform value must surrender effective copyright control through Terms of Service agreements as a condition of participation. In contrast, grassroots platforms consist of cryptographically-identified people operating their networked smartphones independently of any server or global resource; each person holds their own data on their own device, with no third party in possession or intermediation. Here, we define the notion of a digital speech actâa deliberate volitional act by a person of cryptographically signing personal content with the personâs private key, carried out on the personâs own deviceâ through which the person simultaneously establishes attribution, accountability, and authorship over the signed content. We contend that (i) digital speech acts qualify for copyright protection under existing U.S. precedent: Burrow-Giles locates authorship in volitional creative choices despite mechanical or algorithmic processes, Feist supplies the minimal-creativity threshold, and persistent device storage satisfies the Copyright Actâs fixation requirement; (i) the digital social contract underlying grassroots platforms preserves this copyright by designâsigned content cannot be unbundled from its signature, and the full provenance chain accumulates as content is forwardedâso that ownership and possession coalesce in the person; and (i) copyright in digital speech acts is a prerequisite for digital sovereignty and democratic self-governance. * Ehud Shapiro is at the London School of Economics and the Weizmann Institute of Science. The authors used a generative AI tool (Claude by Anthropic) for drafting, restructuring, research and source identification, copy-editing, citation formatting, and L A T E X formatting during the preparation of this manuscript. arXiv:2606.19263v1 [cs.SI] 17 Jun 2026 Table of Contents I Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 I The Current Regime: Platform Governance Through Copyright and Contract . . . 7 A Historical Foundations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 B The Copyright Foundation: Code as Literary Works . . . . . . . . . . . . . . . . . . . 8 C The Copyright Paradox: Software Licensing . . . . . . . . . . . . . . . . . . . . . . . 9 D Juridical Infrastructures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 E From Software Licensing to Platform Terms of Service . . . . . . . . . . . . . . . . . 12 F The Economics of the License Exchange . . . . . . . . . . . . . . . . . . . . . . . . . 12 G Ownership and Possession Separated . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 H The Limitations of Regulatory Responses . . . . . . . . . . . . . . . . . . . . . . . . 14 I Toward Democratic Alternatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 I Grassroots Platforms: Architectural Prevention of Copyright Extraction . . . . . . 15 A Volitional and Machine Transactions . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 B Enforcement via Mutual Attestation . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 C Multiple Independent Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 D No External Legal Entity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 IV Digital Speech Acts: Technical Definition and Legal Foundation . . . . . . . . . . . 16 A Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 B Technical Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 C Technical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 D Two Parallel Foundations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 1Performative Utterances (Searle and Austin) . . . . . . . . . . . . . . . . . . . . . . 18 2Volitional Transactions (VMAT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 E Digital Speech Acts as Legal Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 V Core Legal Framework: Digital Speech Acts as Copyrightable Expression . . . . . . 19 A Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 B What Copyright Protects in Digital Speech Acts . . . . . . . . . . . . . . . . . . . . . 19 C Authorship under Burrow-Giles v. Sarony (1884) . . . . . . . . . . . . . . . . . . . . 20 D Originality under Feist v. Rural (1991) . . . . . . . . . . . . . . . . . . . . . . . . . . 21 E Fixation under 17 U.S.C. § 102(a) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 F Synthesis: Copyright and Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . 23 VI The User/Person Distinction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 A âUserâ as a Legal Category Created by Platform Architecture . . . . . . . . . . . . . . 24 B Personhood in Grassroots Architectures . . . . . . . . . . . . . . . . . . . . . . . . . 25 C Why the Distinction Matters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 VIICopyright Ownership and Digital Sovereignty . . . . . . . . . . . . . . . . . . . . . 26 A Dual Ownership: Cryptographic Possession and Legal Copyright . . . . . . . . . . . . 27 B From Ownership to Sovereignty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 VIIIProvenance, Accountability, and Legal Culpability . . . . . . . . . . . . . . . . . . . 28 A Cryptographic Provenance Chains . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 B Copyright Preserved at Each Step . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 C Accountability and Legal Liability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 1Forwarder Liability: Unresolved Questions . . . . . . . . . . . . . . . . . . . . . . 29 2Accountability and Chilling Effects . . . . . . . . . . . . . . . . . . . . . . . . . . 32 3Accountability Through Provenance . . . . . . . . . . . . . . . . . . . . . . . . . . 32 1 D Evidentiary Authentication of Provenance Chains . . . . . . . . . . . . . . . . . . . . 32 1The Authentication Requirement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 2Self-Authenticating Digital Signatures . . . . . . . . . . . . . . . . . . . . . . . . . 33 3Authentication Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 4Provenance Chain as Business Record . . . . . . . . . . . . . . . . . . . . . . . . . 33 5Challenges and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 6Comparative Advantage Over Platform Evidence . . . . . . . . . . . . . . . . . . . 34 7Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 E International and Jurisdictional Complexity . . . . . . . . . . . . . . . . . . . . . . . 35 1The Limits of a U.S.-Centric Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 35 2Cross-Border Speech and Personal Jurisdiction . . . . . . . . . . . . . . . . . . . . 36 3Choice of Law in Cross-Border Speech . . . . . . . . . . . . . . . . . . . . . . . . 37 4Copyright and the Berne Convention . . . . . . . . . . . . . . . . . . . . . . . . . . 37 5Regulation Without Platform Intermediaries . . . . . . . . . . . . . . . . . . . . . . 38 6Synthesis: Jurisdictional Complexity, Not Simplicity . . . . . . . . . . . . . . . . . 38 IX Addressing Legal Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 A Human Authorship and Algorithmic Execution . . . . . . . . . . . . . . . . . . . . . 39 B Idea, Expression, and Method Under Baker v. Selden . . . . . . . . . . . . . . . . . . 40 C Merger Doctrine and Input-Dictated Expression . . . . . . . . . . . . . . . . . . . . . 42 D Mechanical Reproduction and the Requirement of Originality . . . . . . . . . . . . . . 43 E Thin Copyright, Overclaiming, and the Limits of Protection . . . . . . . . . . . . . . . 45 F Remaining Statutory and Policy Objections . . . . . . . . . . . . . . . . . . . . . . . 46 1Contract Preemption Under Section 301 . . . . . . . . . . . . . . . . . . . . . . . . 46 2First Sale Doctrine and the Limits of Section 109 . . . . . . . . . . . . . . . . . . . 47 3Scholarly Critiques and the Scope of Copyright . . . . . . . . . . . . . . . . . . . . 49 G From Copyrightability to Property and Self-Governance . . . . . . . . . . . . . . . . . 50 X Property and Democratic Self-Governance . . . . . . . . . . . . . . . . . . . . . . . 51 A From Labor to Democratic Independence . . . . . . . . . . . . . . . . . . . . . . . . 51 B Digital Speech Acts as Productive Assets . . . . . . . . . . . . . . . . . . . . . . . . . 52 C Ownership, Possession, and Distributed Power . . . . . . . . . . . . . . . . . . . . . . 53 D Democratic Ownership and Anti-Enclosure . . . . . . . . . . . . . . . . . . . . . . . 54 XI Economic Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 A Grassroots Architectures as Property-Owning Networks . . . . . . . . . . . . . . . . . 55 1Economic Autonomy and Creator Commercialization . . . . . . . . . . . . . . . . . 55 2Intermediary Rent Extraction and Alternative Market Structures . . . . . . . . . . . 56 3Limitations and Open Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 XIIConclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 A What This Establishes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 B Architecture, Ownership, and Power . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 C The Stakes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 D Democratic Alternatives Are Possible . . . . . . . . . . . . . . . . . . . . . . . . . . 59 2 I. INTRODUCTION Platforms wield two legal instruments in tandem: copyright to protect their code, and contract to extract user content. The legal framework enabling this asymmetry emerged through case-by- case litigation over individual software disputes in the 1980s and 1990sâdecisions that prioritized developer interests over user rights, validated adhesion contracts on economic efficiency grounds, and granted comprehensive protection to all forms of computer code. Consider what happens when a person posts on Facebook, uploads to YouTube, or shares on X. Copyright law recognizes the person as the author of their creation, granting exclusive rights under 17 U.S.C. § 106 to reproduce, distribute, and display the work. But to participate in digital lifeâto connect with friends, share with family, engage in public discourseâa person must first agree to Terms of Service granting platforms "worldwide, non-exclusive, transferable, sub-licensable and royalty-free" licenses to everything they create. 1 In return, the person receives revocable access to platform serviceâa license the platform can terminate at any time, for any reason, often without explanation. Meanwhile, platforms assert the full force of copyright law to protect their own code. Every API, every algorithm, every line of machine-readable object code receives comprehensive copyright protection as "literary works." The asymmetry is stark: platforms use copyright defensively over their code and contract offensively over user content. Platform code is protected property; user content is extracted capital. This was not a technological inevitability. The architecture of platform governance emerged through specific historical and technical conditions during the Internetâs formative years. The 1976 Copyright Act extended copyright protection to computer programs while leaving critical questions unresolvedâwhether code in machine-readable form was copyrightable, whether programs stored in RAM were "fixed," and where to draw the line between copyrightable expression and uncopyrightable ideas. 2 Congress delegated these questions to the courts. 3 Federal courts resolved these questions in favor of software developers. Apple Computer v. Franklin Computer (1983) held that all computer codeâeven purely functional operating systems in machine-readable object codeâqualifies for copyright protection. 4 MAI Systems v. Peak Computer (1993) held that even temporary loading of software into RAM creates a "fixed" copy subject to copyright, and that software users are "licensees" bound by all license terms, not "owners" entitled to copyright exemptions. 5 ProCD v. Zeidenberg (1996) held shrinkwrap licenses enforceable under the UCC, reasoning that "competition among vendors, not judicial revision of a packageâs contents, is how consumers are protected in a market economy." 6 These decisions prioritized developer economic interests over user rights. 7 This legal framework developed alongside the Internetâs privatization and commercialization in the late 1980s and early 1990s, within a permissive U.S. regulatory environment that treated Internet service providers as "value-added services" requiring minimal state intervention. 8 Not all courts followed this approachâsubsequent decisions imposed notice and assent requirements that limited enforceabilityâbut platforms adapted their Terms of Service structures to satisfy these standards while preserving the asymmetric license exchange. 9 1 See Meta, Terms of Service (effective Mar. 4, 2026), https://w.facebook.com/terms/ (last visited May 18, 2026); YouTube, Terms of Service (effective Jan. 1, 2025), https://w.youtube.com/static?template=terms (last visited May 18, 2026); X, Terms of Service (effective Nov. 24, 2024), https://x.com/en/tos (last visited May 18, 2026). 2 See infra Part I.B. 3 Id. 4 Apple Computer, Inc. v. Franklin Computer Corp., 714 F.2d 1240, 1249 (3d Cir. 1983). 5 MAI Systems Corp. v. Peak Computer, Inc., 991 F.2d 511, 518 (9th Cir. 1993). 6 ProCD, Inc. v. Zeidenberg, 86 F.3d 1447, 1451 (7th Cir. 1996); see infra Part I.C. 7 See infra Part I. 8 Roxana Radu, Negotiating Internet Governance 61 (2019). 9 See infra Part I.C (discussing Specht v. Netscape, Klocek v. Gateway, Bragg v. Linden Research, and platform adaptation through hybridwrap mechanisms). 3 These decisions established what we term a "copyright paradox" 10 : software licensing simulta- neously evades and depends upon copyright law, creating private governance structures that exceed the intended scope of intellectual property law. 11 Platforms evade copyrightâs limitations by char- acterizing transactions as licenses rather than sales, denying users the exemptions the Copyright Act grants to owners. Yet they depend on copyrightâs enforcement power to validate these licenses as enforceable contracts. The result is a private governance regime through which platforms ob- tain quasi-sovereign control over essential digital communication infrastructuresâa characterization advanced by scholars across antitrust law, information law, and critical platform studies. 12 This legal framework, built upon decisions addressing individual software disputes in a pre- platform world, has scaled to become the architecture which governs billions globally through "hybridwrap" Terms of Service agreements. 13 Every major platform deploys the same structure: comprehensive copyright protection for platform code, mandatory content licenses from users, and revocable access as the only consideration for use. Global regulatory responses, such as the EUâs GDPR and Digital Services Act, have sought to impose protective rules on digital users while leaving the fundamental copyright appropriation structure intact. Platforms are permitted to accumulate concentrated control of user-generated content and data, with regulators then attempting to impose ex post protectionsâprivacy rules, content moderation requirements, data portability rights, and data localization requirements. Yet these regulatory interventions cannot address the underlying issue: platforms retain comprehensive content licenses and physical possession of user data regardless of protective rules or territorial mandates. Regulatory protections address symptoms rather than causes. The cause is the legal and technical architecture which enables copyright appropriation and the separation of ownership from possession. Users may nominally own copyright in their content and data, while platforms receive broad licenses to said content and control the data through centralized servers. Grassroots networks provide a contrasting architectural arrangement. These serverless, dis- tributed, permissionless systems operate solely on the networked smartphones of cryptographically- identified participants and have been formally specified as multiagent systems of volitional transac- tions. 14 In such systems, each person remains in possession of their own data, with no third party in possession or intermediation. 15 Unlike centralized platforms (Facebook, X) or federated systems (Mastodon), grassroots platforms have no external legal entity that can draft Terms of Service or extract comprehensive content licenses from users as a condition of access. Each person stores their data on their own smartphone, with no third party able to obtain access unless explicitly granted. The absence of an operating entity has a direct legal consequence: there is no legal person positioned to require users to grant broad copyright licenses through contractual frameworks as a prerequisite for participation. 16 Within this architectural setting, the central object is the digital speech actâa deliberate vo- litional act by a person, performative in the sense of Searle and Austin, 17 and realized by cryp- tographically signing personal content with the personâs private key. Through this act the person simultaneously establishes attribution (this content is verifiably mine), accountability (I stand behind it and can be held responsible for it), and authorship (I own the resulting expression). 10 James Golike, A Pre-History of Platform Power: A Critical Examination of Early U.S. Legal Discourse Defining Global Platform-User Contractual Relations (2025) (LSE MSc dissertation) (on file with author). 11 See infra Part I. 12 See, e.g., Lina M. Khan, Amazonâs Antitrust Paradox, 126 Yale L.J. 710 (2017) (analyzing platform monopoly power over essential infrastructure); Jack M. Balkin, Information Fiduciaries and the First Amendment, 49 U.C. Davis L. Rev. 1183 (2016) (platforms as exercising quasi-governmental functions); Julie E. Cohen, Between Truth and Power 37â65 (2019) (platforms as infrastructural power); Frank Pasquale, The Black Box Society 19â58 (2015) (platform control resembling sovereign authority). 13 See infra Part I.E. 14 Andy Lewis-Pye & Ehud Shapiro, Volitional Multiagent Atomic Transactions: Describing People and their Machines (2026) (arXiv preprint arXiv:2604.25596). 15 Ehud Shapiro, Grassroots Social Networking: Where People Have Agency over Their Personal Information and Social Graph (2024) (arXiv preprint arXiv:2306.13941). 16 See infra Part I. 17 See infra Part IV.D.1. 4 This Article makes the case that digital speech acts should qualify for copyright protection under existing U.S. precedent. Not through the platform-era authorities that enabled extraction, but through foundational copyright principles predating platforms by decades or centuries. The primary precedent is Burrow-Giles Lithographic Co. v. Sarony (1884), which remains significant not merely because it involved photography, but because it illustrates how copyright law confronts novel technologies. Faced with a medium that mechanically captured images, the Supreme Court did not focus on the cameraâs operation as such. Instead, it asked whether the expressive elements embodied in the resulting work originated with a human author. The Court concluded that authorship existed because the photographer selected the subject, pose, costume, lighting, and arrangement, while the camera merely executed those choices mechanically. 18 Digital speech acts present a similar doctrinal question. The relevant inquiry is not whether cryp- tographic systems participate in the creation of an expression, but whether the expressive choices embodied in the digital speech act originate with the human participant or with the technology itself. A person creating a digital speech act chooses what content to sign, when to sign it, what commu- nicative act to perform, and which cryptographic identity to use. Cryptographic algorithms execute those choices mathematicallyâjust as cameras execute photographersâ choices mechanicallyâbut make no creative decisions themselves. Once authorship is established, Feist Publications, Inc. v. Rural Telephone Service Co. (1991) supplies the governing originality standard. Feist reaffirmed that copyright requires only a minimal degree of creativityâan âextremely lowâ threshold satisfied by modest choices in selection, coordina- tion, and arrangement. 19 The Copyright Actâs fixation requirement likewise applies straightforwardly to digital speech acts stored persistently on participant devices. 20 Taken together, these authorities demonstrate that digital speech acts satisfy the core requirements of copyrightability. Burrow-Giles locates authorship in the human expressive choices embodied in the work, Feist confirms that those choices readily exceed copyrightâs minimal originality threshold, and the Copyright Actâs fixation requirement is satisfied through persistent device storage. Digital speech acts should therefore qualify for copyright protection under established copyright doctrine. Copyright ownership alone, however, is insufficient without architectural enforcement. The digital social contract underlying grassroots networksâa voluntary agreement among people, spec- ified, fulfilled, and enforced through code 21 âspecifies how digital speech acts are handled. It enforces two rules as machine transactions guarded by the presence of a digital speech act: (1) non-unbundlingâsigned content cannot be separated from the signature when forwarded, ensuring the creatorâs copyright claim travels with the content; and (2) provenance preservationâthe full chain of forwarders must be preserved, identifying everyone who forwarded the content. 22 Together, these rules maintain the original copyright and establish clear accountability for both creators and forwarders. The result is the coalescence of ownership and possessionâa fundamental architectural dif- ference from platform regimes. In grassroots architecture the person who owns the copyright in a digital speech act also holds the copyrighted expression on their own device. No third party possesses the content, and no Terms of Service can extract comprehensive licenses from what no third party possesses. Copyright ownership thus becomes practically enforceableânot merely a nominal right surrendered through adhesive licensing, but actual control over reproduction, distribution, and use of oneâs own expressions. The significance of recognizing digital speech acts as copyrightable expression extends beyond copyright doctrine itself. Copyright provides a legal basis for personal control over the creation, circulation, and attribution of expression. When persons both own and possess their digital expres- sions, participation in digital life no longer depends upon third-party custody, platform access, or the 18 Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53, 60 (1884); see infra Part V. 19 Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340, 345 (1991); see infra Part V. 20 17 U.S.C. § 102(a). 21 Luca Cardelli et al., Digital Social Contracts: A Foundation for an Egalitarian and Just Digital Society (2020) (arXiv preprint arXiv:2005.06261). 22 See infra Part IV. 5 surrender of rights through contractual intermediation. The resulting independence suggests a broader conception of digital sovereigntyâthe ability to conduct oneâs digital life free of third-party control, surveillance, manipulation, and rent-seeking. 23 Recognizing copyright in digital speech acts therefore carries broader economic and democratic implications, because it establishes a legal foundation upon which alternative forms of digital orga- nization may be constructed. Rather than treating digital communication primarily as a source of value extraction mediated by platforms, it creates the possibility of organizing digital relationships around the rights and responsibilities of the persons who create and communicate digital expression. Whether such sovereignty ultimately requires forms of property ownership is a question with deep roots in democratic theory. Drawing on traditions associated with Locke, 24 Rawls, 25 and Singer, 26 Part X examines the relationship between property, sovereignty, and democratic participation in digital environments, while engaging substantial criticisms of ownership-based approaches from intellectual property and commons scholarship. The argument that digital speech acts qualify for copyright protection under existing doctrine, however, faces serious objections. Critics may contend that digital speech acts are machine-generated rather than authored, functional rather than expressive, mechanically produced rather than creative, or too thinly original to warrant meaningful protection. Existing doctrine provides support for each of these concerns. Recent algorithmic generation cases (Thaler v. Perlmutter, Naruto v. Slater) deny copyright to machine-created outputs. Idea/expression doctrine (Baker v. Selden) excludes methods from protection. Merger doctrine prevents protection when expression is input-dictated. Mechanical reproduction precedents (Bridgeman Art Library v. Corel Corp., Meshwerks, Inc. v. Toyota Motor Sales U.S.A., Inc.) deny copyright to slavish copies lacking creativity. Thin copyright scholarship warns against minimal creativity enabling strategic overclaiming. Contract preemption (§301), first sale (§109), and copyright skeptics raise doctrinal and policy concerns. Part IX addresses each of these objections, testing the digital speech act framework against the principal doctrinal and policy challenges raised by existing copyright law. The Article proceeds as follows. Part I examines the current regime of platform copyright dominance, tracing how 1980sâ90s copyright and software licensing precedents have scaled into mechanisms for global private governance. Part I describes grassroots platforms and the architec- tural features that preclude third-party possession and intermediation. Part IV defines digital speech acts technically and theoretically, situating them within VMATâs volitional-transaction framework and Searle and Austinâs speech act theory. Part V develops the doctrinal argument, identifying Burrow-Giles as the primary precedent for human authorship despite mechanical execution and drawing additional support from Feistâs originality standard and the Copyright Actâs fixation re- quirement. Part VI develops the distinction between legal persons and platform users, examining how contemporary platforms transform persons into contractually-defined users through Terms of Service while grassroots architectures reestablish authorship, ownership, and accountability through cryptographic identity and binding. Part VII develops the property-sovereignty framework, draw- ing on Locke, Rawls, and Singer to examine the relationship between copyright ownership, digital sovereignty, and democratic participation. Part VIII examines provenance preservation and how it maintains copyright while establishing accountability. Part IX addresses defensive strategies against anticipated legal challengesâalgorithmic generation, idea/expression, mechanical reproduc- tion, merger, thin copyright, contract preemption, first sale, and scholarly critics. Part X explores broader implications for platform governance, content moderation, and democratic digital infras- tructure. Part XI examines the economic implications of the foregoing analysis, considering how architectures that distribute ownership and control to authors may alter the production, exchange, and governance of digital goods and services. Part XII concludes. This Articleâs central contribution is demonstrating that foundational copyright principles, ap- 23 Ehud Shapiro, Grassroots Systems: Concept, Examples, Implementation and Applications, in Proceedings of the 37th International Symposium on Distributed Computing (DISC 2023) 47:1â47:18 (2023) (also at arXiv:2301.04391). 24 John Locke, Second Treatise of Government § 25â51 (C.B. Macpherson ed., 1980) (1690). 25 John Rawls, Justice as Fairness: A Restatement 135â40 (Erin Kelly ed., 2001). 26 Joseph William Singer, Entitlement: The Paradoxes of Property (2000). 6 plied within grassroots architectures, can support person-centered ownership rather than platform extraction. The argument does not depend on legal reform. Instead, it argues that existing copyright doctrine may produce different ownership outcomes when expression is created, possessed, and circulated within architectures that do not rely on third-party custody or contractual intermediation. If digital speech acts qualify for copyright protection, then copyright law may provide more than a mechanism for protecting software and platform business models; it may also furnish a legal founda- tion for alternative forms of digital organization grounded in distributed ownership, accountability, and democratic participation. I. THE CURRENT REGIME: PLATFORM GOVERNANCE THROUGH COPYRIGHT AND CONTRACT A. Historical Foundations The dominance of U.S.-based digital platforms over global Internet architecture is not happen- stance, but the product of specific historical, regulatory, and legal conditions during the Internetâs formative years. Under an economic paradigm termed "collective invention," the early development of the Internet took place almost entirely within the United States. 27 Essential protocols, including file transfer and TCP/IP, were developed in the early 1970s while the Internet (then "ARPANET") was still a research project owned by the U.S. Department of Defense. 28 The principle of "end-to-end" designâplacing computing power on usersâ devices rather than in network infrastructureâwas articulated by Saltzer, Reed, and Clark in 1984 29 and played an integral role during the Internetâs stewardship by the U.S. National Science Foundation from 1985 to 1995. 30 As this design principle guided the networkâs technical architecture, privatizationâthe transfer of publicly-owned Internet infrastructure to private actors who marketed and sold access to usersâbegan in the late 1980s and accelerated through the early 1990s, transforming the nascent Internet from a U.S. national communication network to the dominant global network of data exchange it is today. 31 This privatization occurred within a distinctive regulatory environment. While the U.S. gov- ernment had provided intensive funding and research support during the Internetâs development as a public infrastructure, its approach shifted as privatization progressed. The government in- creasingly adopted permissive regulatory and policy approaches designed to promote the Internetâs commercial proliferation, reducing direct state intervention and giving preferential treatment to market mechanisms. 32 In 1988, the Federal Communications Commission created the category of "value-added services" for digital communications, leaving computer-mediated information virtually unregulated. 33 The Clinton Administrationâs 1996 "Framework for Global Electronic Commerce" ce- mented this approach, stressing self-regulation and private sector development of Internet operations to governments around the world. 34 Within this unique context, U.S. courts emerged at critical junctures in the Internetâs development. While early technical standards and governance structures arose from quasi-private arrangements, the 27 Shane Greenstein, How the Internet Became Commercial: Innovation, Privatization, and the Birth of a New Network (2015). 28 Radu, supra note 8, at 44. 29 J.H. Saltzer, D.P. Reed & D.D. Clark, End-to-End Arguments in System Design, 2 ACM Transactions on Computer Systems 277 (1984). 30 Greenstein, supra note 27. 31 Julia Pohle & Daniel Voelsen, Centrality and Power: The Struggle over the Techno-Political Configuration of the Internet and the Global Digital Order, 14 Polây & Internet 13 (2022). 32 Radu, supra note 8. 33 Amendment of Section 64.702 of the Commissionâs Rules and Regulations (Third Computer Inquiry), Report and Order, 104 F.C.C.2d 958 (1986), reconsidered, 2 F.C.C. Rcd. 3035 (1987), further reconsidered, 3 F.C.C. Rcd. 1135 (1988); see also Radu, supra note 8, at 45. 34 William J. Clinton & Albert Gore, Jr., A Framework for Global Electronic Commerce (1997), available at https://clintonwhitehouse4.archives.gov/WH/New/Commerce/ 7 stateâs deliberate privatization and multi-stakeholder governance approach created an environment where private actors sought judicial mediation only on issues that could not be resolved privately. Federal courts became the forum for resolving disputes over domains, boundaries, rights, and obligations of private actors within Internet architectureâand, as Part I demonstrates, consistently resolved ambiguities in copyright law in ways that favored software developers over users. 35 B. The Copyright Foundation: Code as Literary Works The legal foundation for platform copyright dominance begins with the Copyright Act of 1976 and its 1980 amendments addressing computer programs. Congress included computer programs within the definition of "literary work"â"works, other than audiovisual works, expressed in words, numbers, or other verbal or numerical symbols or indicia"âbut only "to the extent that they incor- porate authorship in the programmerâs expression of original ideas as distinguished from the ideas themselves." 36 This qualified inclusion, however, left critical questions unresolved. Many aspects unique to computer code remained unaddressed: whether a program contained entirely within a computerâs electronics was a "copy," whether temporary storage in a computer constituted making a "copy," and whether copyright law protects work that blends expression and utility. 37 To address these concerns, Congress created the National Commission on New Technological Uses of Copyrighted Works (CONTU) in 1974, which reviewed whether computer programs met copyrightability requirements and whether programs in machine-readable form were copyrightable material. 38 CONTUâs report, published in 1978 after the passage of the 1976 Copyright Act, led to recommendations that Congress adopted substantially in the 1980 amendments. 39 Critically, CONTU recognized it was impossible to draw a clear line between copyrightable "expression" and uncopyrightable "ideas" in computer programsâa distinction central to copyright law. 40 Drawing the line too far toward copyright protection would grant strong monopolies over specific applica- tions; drawing it too far toward public domain would allow easy copying of programmersâ work, discouraging creation. 41 Unable to resolve this tension, CONTU and Congress left this fundamental question to the courts. The Third Circuitâs decision in Apple Computer v. Franklin Computer (1983) provided the answer: comprehensive copyright protection for all computer code, regardless of function, form, or medium. Franklin Computer copied portions of Appleâs operating system software to create compatible computers, defending on grounds that purely functional codeâespecially operating systems in machine-readable object code stored in ROMâcould not be copyrighted. 42 The Third Circuit rejected every aspect of this defense. Against the argument that object code (binary sequences of 0s and 1s unintelligible to humans) was uncopyrightable due to unintelligibility, the court held it was "the clear intention of Congress to protect all computer code." 43 Against the argument that code was uncopyrightable because it had utilitarian purpose, the court held that Congress made "no distinction between application programs and operating programs." 44 Against the argument that code stored in ROM was a mechanical device rather than a fixed expression, the court held that "fixation... is satisfied through the embodiment of the expression in the ROM devices." 45 35 See infra Part I.BâC. 36 Lajos P. Pataki, Jr., Copyright Protection for Computer Programs Under the 1976 Copyright Act, 52 Ind. L.J. 503, 504-05 (1977). 37 Id. 38 National Commission on New Technological Uses of Copyrighted Works, Final Report of the National Commission on New Technological Uses of Copyrighted Works 9 (1978). 39 Pamela Samuelson, CONTU Revisited: The Case Against Copyright Protection for Computer Programs in Machine-Readable Form, 1984 Duke L.J. 663 (1984). 40 CONTU Final Report, supra note 38, at 22. 41 Peter S. Menell, An Analysis of the Scope of Copyright Protection for Application Programs, 45 Stan. L. Rev. 1045 (1993). 42 Apple Computer, Inc. v. Franklin Computer Corp., 545 F. Supp. 812 (E.D. Pa. 1982). 43 Apple Computer, Inc. v. Franklin Computer Corp., 714 F.2d 1240, 1249 (3d Cir. 1983). 44 Id. 45 Id. 8 The decision established comprehensive copyright protection for computer code and resolved critical questions Congress had left open. The Third Circuit held that all computer code receives copyright protection regardless of formâsource code or object codeâmedium of storageâROM chips, semiconductor devices, or magnetic disksâor functionâapplication programs or operating systems. 46 By rejecting Franklinâs arguments that object code was too unintelligible for copyright, that code stored in ROM was a utilitarian object rather than fixed expression, and that operating systems were purely functional works outside copyrightâs scope, the court granted software manu- facturers extensive exclusive rights over their entire code base. Subsequent litigation refined the boundaries of this protection without challenging its core. Courts addressed whether reverse engineering for interoperability constituted fair use, 47 whether menu command hierarchies were copyrightable expression or uncopyrightable methods of operation, 48 and where precisely to draw the idea/expression line for functional program elements. But the fundamental holding that all computer code qualifies for copyright protection regardless of form or function has remained settled law. C. The Copyright Paradox: Software Licensing Even before Apple v. Franklin established comprehensive code protection, the software industry had intentionally developed a mechanism in the 1970s to circumvent copyright exemptions: the software license. 49 Concerned that competitors could purchase software copies and then lease them to consumers (who could duplicate them) or prepare derivative programs, software developers deliberately characterized transactions as licenses rather than sales. 50 By making each license "personal and non-transferable," producers specifically targeted the first-sale doctrineâthe copyright exemption allowing rightful owners of copyrighted material to sell, lend, or destroy the physical object without the copyright holderâs consent. 51 This created the copyright paradox. Software licenses serve dual purposes: controlling the number of users, permitted locations, uses, and duration of software use, 52 while simultaneously transforming the enforcement mechanism for unauthorized use. By characterizing transactions as licenses rather than sales, any breach of license terms becomes a contract law violation rather than copyright infringement. 53 Copyright law validates the license as an enforceable contract, 54 but the license itself disenfranchises users from the protections copyright law providesâa structure that would later scale from individual software transactions to global platform governance. 55 The paradox was reinforced by MAI Systems v. Peak Computer (1993), which established two critical precedents. First, the Ninth Circuit held that merely loading software into RAMâthe most temporary form of computer storageâconstitutes creating a "fixed" copy subject to copyright pro- tection. 56 The courtâs reasoning was widely criticized by legal scholars as illogical: it acknowledged that RAM copies are "destroyed when a computer is turned off" yet held them sufficiently permanent to constitute copies under the Copyright Act. 57 Although Congress later amended Section 117 to 46 Jeffrey L. Nussbaum, Apple Computer, Inc. v. Franklin Computer Corporation Puts the Byte Back into Copyright Protection for Computer Programs, 14 Golden Gate U. L. Rev. 281, 281â82 (1984). 47 Sega Enters. Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1992). 48 Lotus Dev. Corp. v. Borland Intâl, Inc., 49 F.3d 807 (1st Cir. 1995), affâd by an equally divided court, 516 U.S. 233 (1996). 49 Michael L. Rustad, Software Licensing: Principles and Practical Strategies, Suffolk U. L. Sch. Legal Studies Research Paper Series, Research Paper 14-5 (2014). 50 Id. 51 Id. 52 Rustad, supra note 49. 53 Nancy S. Kim, The Software Licensing Dilemma, 2008 BYU L. Rev. 1103 (2008). 54 Michael J. Madison, Reconstructing the Software License, 35 Loy. U. Chi. L.J. 275 (2003). 55 Id. 56 MAI Systems Corp. v. Peak Computer, Inc., 991 F.2d 511, 518 (9th Cir. 1993). 57 Kathryn Levin, Intellectual Property Law - MAI v. Peak: Should Loading Operating System Software into RAM Constitute Copyright Infringement?, 24 Golden Gate U. L. Rev. 649 (1994); Bridget J. Murphy, Loading Software into RAM Creates a Copy: MAI Systems Corp. v. Peak Computer Inc., 10 Santa Clara High Tech. 9 permit temporary RAM copies for maintenance and repair purposes, 58 the core holding that RAM loading constitutes fixation has remained. This extension of copyright protection granted software manufacturers "an almost patent-like monopoly over their computer code and software programs," extending exclusive rights at the expense of software users. 59 Second, the court validated software licensing as a viable framework for software transactions. MAI needed only to prove copyright ownership over their code; the license agreement then controlled the relationship, characterizing software customers as "licensees" bound by all license terms, without the Section 117 exemptions granted to "owners." 60 This distinctionâlicensee versus ownerâbecame central to platform power: users could be denied the rights that copyright law grants to owners simply through contractual characterization. The final component of platform copyright dominance emerged with ProCD v. Zeidenberg (1996), which validated "shrinkwrap" licensesâcontracts enclosed within product packaging that become binding when users open the package. The Third Circuit had previously rejected such licenses in Step-Saver Data Systems v. Wyse Technology, holding that terms presented after sale could not bind purchasers. 61 Matthew Zeidenberg purchased ProCDâs SelectPhone software at retail, ignored the enclosed license restricting use to non-commercial purposes, and resold the database over the Internet. 62 ProCD sued for breach of license terms. 63 Zeidenberg defended on the ground that the license terms were enclosed inside the package and therefore could not bind him at the moment of purchaseâthe same reasoning the Third Circuit had accepted in Step-Saver. The Seventh Circuit, in an opinion by Judge Frank Easterbrook, held that shrinkwrap licenses are enforceable contracts under the Uniform Commercial Code, directly contradicting Step-Saver. 64 Judge Easterbrook analogized shrinkwrap licenses to other standard-form contracts such as insurance policies and airline tickets, where "the exchange of money precedes the communication of detailed terms." 65 Merely providing the opportunity to read license terms was sufficient for enforceability: "Notice on the outside, terms on the inside, and a right to return the software for a refund if the terms are unacceptable... may be a means of doing business valuable to buyers and sellers alike." 66 Dismissing concerns about unequal bargaining power, Judge Easterbrook wrote: "Competition among vendors, not judicial revision of a packageâs contents, is how consumers are protected in a market economy." 67 The court justified this conclusion on economic efficiency grounds: standard form contracts saved both consumers and producers time and money in negotiating individual contracts, and software consumers were "better off" under enforceable shrinkwrap licenses. 68 ProCD created a circuit split that opened the legal market for adhesion contract experimentation. With courts divided on enforceability, companies could draft Terms of Service and test their limits, knowing some jurisdictions would uphold them. Subsequent litigation did not overturn ProCDâs core holding that online adhesion contracts can be enforceable. Instead, courts refined the boundaries of enforceability. The Second Circuit held in Specht v. Netscape that browsewrap agreements fail when users lack sufficient notice of terms. 69 The District of Kansas held in Klocek v. Gateway that terms inside a box are not part of the contract when the contract is formed at purchase. 70 L.J. 499 (1994). 58 17 U.S.C. § 117(c) (1998). 59 Theodore Arriola, Software Copyright Infringement Claims After MAI Systems v. Peak Computer, 69 Wash. L. Rev. 405 (1994). 60 Michael E. Johnson, The Uncertain Future of Computer Software Usersâ Rights in the Aftermath of MAI Systems, 44 Duke L.J. 327 (1994). 61 Step-Saver Data Sys., Inc. v. Wyse Tech., 939 F.2d 91 (3d Cir. 1991). 62 ProCD, supra note 6 63 Id. 64 Id. at 1450. 65 Id. at 1449. 66 Id. 67 Id. at 1451. 68 Brett L. Grusar, Contracting Beyond Copyright: ProCD, Inc. v. Zeidenberg, 10 Harv. J.L. & Tech. 353 (1997). 69 Specht v. Netscape Commcâns Corp., 306 F.3d 17, 35 (2d Cir. 2002). 70 Klocek v. Gateway, Inc., 104 F. Supp. 2d 1332, 1341 (D. Kan. 2000). 10 The Eastern District of Pennsylvania in Bragg v. Linden Research found certain Terms of Service unconscionable. 71 These decisions established limiting principlesâconspicuous notice, meaningful opportunity to review, substantive unconscionability reviewâwithout rejecting adhesion contracts altogether. Platforms adapted by implementing mechanisms combining clickwrap (requiring explicit "I agree" clicks) and browsewrap (presenting terms via conspicuous hyperlinks), adding arbitration clauses to avoid judicial review of substantive terms, and structuring agreements to satisfy the notice require- ments courts demanded. 72 D. Juridical Infrastructures The significance of Apple, MAI, and ProCD extends beyond the specific doctrines each case established. Taken together, these decisions created a legal architecture that enabled new forms of governance through software and contracts. The cumulative effect is best understood as the construction of what this Article terms juridical infrastructures. 73 This Article conceptualizes juridical infrastructures as legal frameworks that function as un- derlying structures through which governance is exercised. The concept combines insights from infrastructure studies and critical legal scholarship. Infrastructure scholars have demonstrated that infrastructures are not merely technical systems but socio-technical arrangements that enable and constrain particular forms of action while embedding relationships of power within seemingly neu- tral systems. 74 Critical legal scholarship similarly observes that law does not merely regulate social relations but constitutes the conditions under which particular social, economic, and political ar- rangements become possible. 75 Juridical infrastructures combine these insights. Just as technical infrastructures shape what actions can occur within a system, legal arrangements allocate rights, obligations, and authority, determining who may participate, on what terms, and subject to whose control. Copyright doctrines, licensing arrangements, contractual forms, and judicial precedents are therefore not merely rules governing platforms. They are part of the operative infrastructure through which platform governance itself becomes possible. The interaction of Apple, MAI, and ProCD illustrates this process. Apple established broad copyright protection for software. MAI transformed software use into licensed access rather than ownership. ProCD validated contractual restrictions imposed through mass-market licenses. None of these decisions created platform governance independently. Together, however, they generated the legal conditions that allowed platform firms to govern users through proprietary software, licensing arrangements, and contractual terms. The result preserved the asymmetric license exchange validated in ProCD while satisfying the boundaries subsequent courts established. As with Appleâs comprehensive protection of software and MAIâs license-centered conception of use, the core precedents remained intact even as courts refined their boundaries through later litigation. What emerged was not merely a collection of doctrines but a durable juridical infrastructure capable of supporting platform governance at global scale. The significance of juridical infrastructures is not merely explanatory. Once established, they become operationalized through concrete institutional arrangements. In the platform economy, the principal mechanism through which these infrastructures are deployed is the Terms of Service agreement. 71 Bragg v. Linden Research, Inc., 487 F. Supp. 2d 593, 607 (E.D. Pa. 2007). 72 See Nancy Kim, Wrap Contracts: Foundations and Ramifications (2013). 73 The concept of âjuridical infrastructuresâ is developed more fully in James Golike, The Copyright Paradox: How Three Court Cases Created Juridical Infrastructures for Platform Governance (on file with author). 74 See Susan Leigh Star, The Ethnography of Infrastructure, 43 Am. Behav. Scientist 377 (1999); Jean- Christophe Plantin et al., Infrastructure Studies Meet Platform Studies in the Age of Google and Facebook, 20 New Media & Socây 293 (2018). 75 See Julie E. Cohen, Between Truth and Power: The Legal Constructions of Informational Capitalism (2019); Alan Hunt, The Concept of Law and the Everyday World, in Reading Dworkin Critically 35 (1992). 11 E. From Software Licensing to Platform Terms of Service The legal framework established by Apple v. Franklin (comprehensive code protection), MAI v. Peak (licensee status denies ownership rights), and ProCD (shrinkwrap validation on efficiency grounds and opportunity-to-read sufficiency) scaled from individual software transactions to global platform governance through "hybridwrap" Terms of Service agreements. When software tran- sitioned from physical retail packages to Internet downloads and web-based services, shrinkwrap licenses evolved into "clickwrap" (requiring users to click "I agree") and "browsewrap" (presenting terms via hyperlink). 76 Contemporary platform Terms of Service combine these approaches: users encounter a hyperlink to Terms of Service upon signup, and any use of the platform constitutes acceptanceâthe hyperlink providing the "opportunity to read" that ProCD deemed sufficient for enforceability. 77 Every major platform follows this approach. Facebookâs Terms of Service state: "If you do not agree to these Terms, then do not access or use Facebook or the other products and services covered by these Terms." 78 YouTube provides: "If you do not understand the Agreement, or do not accept any part of it, then you may not use the Service." 79 X declares: "By using the Services you agree to be bound by these Terms." 80 These agreements are offered in pure "take-it-or-leave-it" structure: users agree by entering the platform, regardless of whether they have read or understood any terms. 81 At the core of every platform Terms of Service is a license exchange that operationalizes the copyright paradox at global scale. Platforms grant users a "personal, worldwide, royalty-free, non- assignable and non-exclusive license to use the software" provided by the platform. 82 This license is revocableâplatforms can terminate user accounts at any time, for any reason, often without explanation. In exchange, users grant platforms comprehensive licenses to user-generated content. Facebook requires users to grant a "non-exclusive, transferable, sub-licensable, royalty-free and worldwide licence to host, use, distribute, modify, run, copy, publicly perform or display, translate and create derivative works of" all user content. 83 YouTube requires a "worldwide, non-exclusive, royalty-free, transferable, sublicensable licence to use that Content (including to reproduce, distribute, modify, display and perform it)" for operating, promoting, and improving the service, plus the "right to monetize" user content. 84 X requires a "worldwide, non-exclusive, royalty-free license (with the right to sublicense) to use, copy, reproduce, process, adapt, modify, publish, transmit, display, upload, download, and distribute" user content "in any and all media or distribution methods now known or later developed, for any purpose," including to "analyze text and other information" and let others do the same, "with no compensation paid." 85 F. The Economics of the License Exchange Platforms assert comprehensive copyright protection over every line of code in their ecosys- temsâfrom operating systems to APIsâthrough the precedents of Apple v. Franklin. Users are characterized as "licensees" rather than "owners" through the precedent of MAI v. Peak, denying them the rights copyright law grants to owners. The license exchange is validated through the precedent of ProCD, despite the complete absence of competitive markets that supposedly justify adhesion contracts. Users nominally retain copyright in content they create, but platforms obtain such broad licensesâworldwide, sublicensable, royalty-free, for any purpose including commercial use and AI trainingâthat nominal ownership becomes practically meaningless. X makes the extraction explicit: 76 Michael Karanicolas, Too Long; Didnât Read: Finding Meaning in Platformsâ Terms of Service Agreements, SSRN Working Paper No. 3887753 (2021). 77 Id. 78 Meta, Terms of Service, supra note 1 79 YouTube, Terms of Service, supra note 1. 80 X, Terms of Service, supra note 1. 81 Grusar, supra note 68. 82 YouTube, Terms of Service, supra note 1 83 Meta, Terms of Service, supra note 1 84 YouTube, Terms of Service, supra note 1 85 X, Terms of Service, supra note 1 12 "the use of the Services by you is hereby agreed as being sufficient compensation for the Content and grant of rights herein." 86 Platform access itself becomes payment, transforming user expressions into platform capital without direct monetary exchange. YouTubeâs Terms state bluntly: "This Agreement does not entitle you to any payments" from monetization of user content. 87 This contractual structure enables platforms to accumulate vast wealth extracted from user- generated value. Meta reported revenues of $200.97 billion in 2025, generating âsubstantially allâ of its revenue from advertising based on user data and content. 88 Alphabet (Google/YouTube) generated $402.8 billion the same year, with âmore than 70%â from online advertising systems that similarly monetize user-generated content, searches, and engagement data without compensation to users. 89 This extraction occurs through Terms of Service agreements that grant platforms comprehensive control over access, use, and interaction within digital spaces while compelling users to surrender licensing rights to the data and content that generates this wealth. The contractual appropriation extends beyond the core license exchange, as the terms and con- ditions outlined in platform Terms of Service demonstrate continuing erosion of user protections. Facebookâs Terms significantly omit any explicit software license grant to users, suggesting a shift from licensing contracts toward pure property-based governance where platform access becomes revocable privilege rather than contractual right. 90 The variation in notification periods for Terms of Service changes reveals graduated erosion: Facebook commits to "at least 30 days" notice before changes; YouTube offers only "reasonable advance notice"; X merely states "We will try to notify you of material revisions." 91 Users have no ability to negotiate, debate, or refuse changesâcontinued use constitutes acceptance. What emerged as solutions to individual software licensing disputes has scaled into comprehensive mechanisms for private governance of global digital communication and commerce. G. Ownership and Possession Separated Beyond the contractual asymmetry, the platform architecture creates a second asymmetry: the separation of ownership and possession. Users in principle own copyright in the content they create, but the centralized server-based infrastructure ensures platforms physically retain the copyrighted expression. The person holds legal rights under 17 U.S.C. § 106; the platform holds the data and controls access through both Terms of Service and platform technical infrastructure. The platform exploits this separation to exercise what legal scholars have characterized as infras- tructural power over users and their data. 92 Platforms obtain personal information, social graph data, and communication content; commercially exploit them through what Zuboff terms "surveillance capitalism"; 93 and render user copyright ownership effectively meaningless because access to oneâs own data depends on platform discretion. This separation of ownership from possessionâusers retain legal rights while platforms exercise physical and technical controlâenables the extraction that Terms of Service agreements legitimate. 86 X, Terms of Service, supra note 1 87 YouTube, Terms of Service, supra note 1 88 Meta Platforms, Inc., Annual Report (Form 10-K) for fiscal year ended Dec. 31, 2025, at 15, 60 (filed Jan. 30, 2026), https://w.sec.gov/ix?doc=/Archives/edgar/data/0001326801/000162828026003942/meta- 20251231.htm [hereinafter Meta 2025 Form 10-K]. 89 Alphabet Inc., Annual Report (Form 10-K) for fiscal year ended Dec. 31, 2025, at 9, 32 (filed Jan. 31, 2026), https://w.sec.gov/ix?doc=/Archives/edgar/data/0001652044/000165204426000018/goog-20251231.htm. 90 Meta, Terms of Service, supra note 1. 91 Meta, Terms of Service § 4.1, supra note 1; YouTube, Terms of Service, supra note 1; X, Terms of Service, supra note 1. 92 See Julie E. Cohen, Configuring the Networked Self: Law, Code, and the Play of Everyday Practice 122â52 (2012) (analyzing how platform architecture separates user ownership from platform control); Balkin, supra note 12, at 1186â89 (platforms as fiduciaries controlling user information); Pasquale, supra note 12, at 19â58 (opacity of platform control over user data). 93 Shoshana Zuboff, The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power (2019). 13 H. The Limitations of Regulatory Responses Regulatory responses to platform dominance have addressed data protection and content gov- ernance without challenging the underlying contractual and technical framework that enables ex- traction. The European Unionâs GDPR imposes consent requirements, data portability rights, and privacy protections. 94 The Digital Services Act requires content moderation transparency and ap- peals mechanisms. 95 These regulatory interventions have achieved meaningful protections: GDPRâs consent requirements and right to erasure provide users some control over personal data process- ing, 96 and the Digital Services Actâs transparency obligations improve accountability for content moderation decisions. 97 Yet these protections leave the fundamental copyright appropriation structure intact. Platforms continue to require users to grant comprehensive content licenses as a condition of access. Users continue to surrender possession of their data and content to platform servers. Platforms continue to extract commercial value from user-generated content without compensationâGDPR does not prevent Meta from requiring users to grant "worldwide, non-exclusive, transferable, sub-licensable, royalty-free" licenses to everything they create, nor does the Digital Services Act prevent platforms from monetizing user content through advertising. Privacy protections and content moderation rules address governance of data already surrendered, not the initial transfer of possession and licensing of rights. This Articleâs argument is not that GDPR and the Digital Services Act accomplish nothing, but that they cannot address the power asymmetry this Article identifies: the separation of ownership from possession enabled by copyright law and operationalized through platform architecture. Regulatory protections operate downstream from the moment of appropriationâthey govern what platforms may do with data and content users have already been compelled to surrender. The architectural intervention this Article proposes operates upstream: preventing the separation of ownership and possession by design, making regulatory protection of already-surrendered data unnecessary. I. Toward Democratic Alternatives The genealogical analysis reveals that current platform dominance is historically constructed rather than technologically determined. The copyright and licensing precedents of the 1980s- 90sâdecided on individual software disputes in a pre-platform eraâscaled into mechanisms for global private governance not through inevitable technological evolution but through specific legal and architectural choices. Platforms employ copyright to protect their code and to legally validate software licensing transactions, characterize users as licensees without ownership rights, and impose adhesion contracts validated on economic assumptions that have proved demonstrably false in plat- form markets. The result is quasi-sovereign corporate control over essential digital infrastructure, exercised through Terms of Service that appropriate user expressions while providing only revocable access to platform services. This historical construction, however, reveals possibilities for democratic alternatives. If platform power rests on specific legal precedents applied through centralized architectures, then the same legal system applied through different architectures can produce the different results. This Article develops an alternative: grassroots architectures that prevent appropriation by design. These architectures operate through digital speech actsâcryptographically signed expressions that persons create and store on their own devices. Because persons both create and possess these 94 Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data (General Data Protection Regulation). 95 Regulation (EU) 2022/2065 of the European Parliament and of the Council of 19 October 2022 on a Single Market For Digital Services (Digital Services Act). 96 See Daniel J. Solove & Paul M. Schwartz, Privacy Law Fundamentals 115â47 (2021) (analyzing GDPRâs effectiveness in constraining data processing). 97 See Daphne Keller, Empirical Evidence of Over-Removal by Internet Companies Under Intermediary Liability Laws, Stanford Ctr. for Internet & Socây (2015) (analyzing platform moderation under various regulatory regimes). 14 copyrighted expressions, ownership and possession coalesceâestablishing person ownership rather than platform extraction. The current regime of platform copyright dominance is neither natural nor inevitable. It is the product of specific historical conditions, legal decisions, and architectural choicesâwhich means it can be reconstructed through different architectural choices that prioritize person agency over corporate extraction. I. GRASSROOTS PLATFORMS: ARCHITECTURAL PREVENTION OF COPYRIGHT EXTRACTION This section describes the grassroots architecture: serverless, permissionless platforms that operate solely on the networked smartphones of their cryptographically-identified participants, in which each person retains ownership and control over their personal information. 98 A. Volitional and Machine Transactions The formal framework for grassroots platforms uses volitional multiagent atomic transactions. 99 Each agentâs state has two components: volitionsâwhat the person wantsâand machine stateâthe state of their device. Transactions are of two kinds: volition transactions, in which a single person freely changes their volitions, and machine transactions, which change only machine state and are guarded by conditions on the volitions of the participating agents. Creating a digital speech actâdeliberately signing contentâis a volition transaction: a free, unary act by a single person, with no precondition. The person decides to sign specific content; this decision is theirs alone. The digital social contract then specifies machine transactions that are guarded by the presence of a digital speech act. Two such rules are central to copyright: Non-unbundling. When a digital speech act is forwarded, the signed content cannot be separated from the signature. The digital social contract enforces this as a machine transaction: forwarding copyrighted content requires embedding the original digital speech actâsignature and allâwithin the forwarderâs new digital speech act. A recipient who receives copyrighted content receives the creatorâs signature with it, and cannot strip it. Provenance preservation. When a digital speech act is forwarded, the full provenance chainâ the sequence of forwardersâmust be included. Each forwarderâs signature is added to the chain, ensuring that every person who forwarded the content is identified, establishing accountability at each step. Together, non-unbundling and provenance preservation are the two rules by which the digital social contract enforces copyright: the creatorâs claim cannot be stripped, and the full history of who forwarded the content is maintained. B. Enforcement via Mutual Attestation A digital social contract must be enforced: participants must not be able to deviate from the contract as programmed. For grassroots platforms, enforcement is achieved through mutual attes- tation: when two agents communicate, each proves to the other that the message was produced by a correct execution of the agreed-upon contract code. An attested agent either executes correctly or crashesâit cannot behave arbitrarily. 100 For digital social contracts that require no consensusâincluding social graphs and social networkingâmutual attestation among the participants suffices for enforcement. The contract runs 98 Shapiro, supra note 23. 99 Lewis-Pye & Shapiro, supra note 14. 100 Ehud Shapiro, Child-Safe Social Networking: Design, Specification, and Implementation (2026) (unpublished manuscript) (on file with author); see also Leslie Lamport, Robert Shostak & Marshall Pease, The Byzantine Generals Problem, 4 ACM Transactions on Programming Languages and Systems 382 (1982). 15 on the participantsâ own devices, and compliance is verified bilaterally. No third-party validators, no global consensus protocol, and no blockchain are required. 101 C. Multiple Independent Instances A grassroots platform can have multiple instances that emerge and operate independently of each other and of any global resource except the network, and can interoperate and coalesce once interconnected, potentially forming a single global instance. 102 This is in contrast to global platforms, which can only have a single instanceâone Facebook, one Bitcoinâand to federated systems such as Mastodon, where each instance depends on its operatorâs server. 103 The grassroots property ensures that any group of people with smartphones can form their own platform instance, independently of any existing instance or global resource. Independently-formed instances may interconnectâfor example, when a person in one instance befriends a person in anotherâand may eventually coalesce into a larger instance. At no point does any third party obtain control over the platform or the digital speech acts of its participants. D. No External Legal Entity In a grassroots platform, there are no servers, no platform operators, no data collection, no algorithmic content selection, and no advertising. 104 Each person stores only the local neighbourhood pertaining to them on their own smartphone, and no third party has access unless explicitly granted. The absence of an operating entity has a direct legal consequence: there is no legal person that can draft Terms of Service, impose license agreements, or claim copyright ownership over the expressions of the platformâs participants. On existing platforms, the entity that operates the platform is also the entity that imposes Terms of Service requiring people to grant broad copyright licenses as a condition of participation. In a grassroots platform, no such entity exists. IV. DIGITAL SPEECH ACTS: TECHNICAL DEFINITION AND LEGAL FOUNDATION A. Overview Cardelli et al. introduced the term âdigital speech actâ to describe a cryptographically-signed, indexed transaction (an âutteranceâ) within a digital social contractâa voluntary agreement among a set of people, specified, fulfilled, and enforced digitally. 105 In its original formulation, a digital speech act by an agent consists of signing an utterance with the agentâs private key and broadcasting it to the other parties of the contract. Cardelli et al. focus on digital speech acts as primitives for constructing egalitarian social contracts through computational mechanisms. This Article extends the concept in three directions relevant to digital self-governance. First, we ground digital speech acts in speech act theory (Searle and Austin) and volitional multiagent atomic transactions (VMAT), 106 establishing that digital speech acts are deliberate performative utterances, not merely cryptographic operations. Second, we demonstrate that digital speech acts should qualify for copyright protection under existing U.S. precedent, 107 providing legal enforceability beyond the computational guarantees Cardelli et al. analyze. Third, we connect the legal status of digital speech acts, including the copyright interests they may generate, to political theory on property and 101 Shapiro, supra note 100. 102 Ehud Shapiro, Grassroots Platforms with Atomic Transactions: Social Networks, Cryptocurrencies, and Democratic Federations, in Proceedings of the 27th International Conference on Distributed Computing and Networking (ICDCN 2026) 71â81 (2026) (also at arXiv:2502.11299). 103 Aravindh Raman et al., Challenges in the Decentralised Web: The Mastodon Case, in Proceedings of the Internet Measurement Conference 217â222 (2019). 104 Shapiro, supra note 15. 105 Cardelli et al., supra note 21. 106 See infra Part IV.D. 107 See infra Part V. 16 democratic sovereignty, 108 establishing that recognition of such rights is an important prerequisite for digital self-governance. These extensions transform digital speech acts from computational primitives into legal objects with both technical and normative force. A digital speech act, as developed here, is a deliberate volitional act by a person, performative in the sense of Searle and Austin, realized by cryptographically signing personal contentâtext, audio, photo, video, or structured dataâwith the personâs private key. By performing this act, the person simultaneously establishes (i) attributionâthe content is verifiably theirs; (i) accountabilityâthey stand behind it and can be held responsible for it; and (i) copyright over the signed expression under applicable law. A digital speech act exists independently of any particular social contract or platform; it is an act of a person, carried out on the personâs own device. Digital speech acts arise in many contexts: transactions in grassroots currencies and bonds 109 are digital speech acts, as is voting in democratic digital communities. 110 Here, we focus on digital speech acts in the context of grassroots social networks. B. Technical Definition A digital speech act is created through the combination of three elements: a self-chosen cryptographic key pair that uniquely identifies the person, content created or selected by the person (text, audio, photo, video, or structured data), and the cryptographic signature produced by signing that content with the personâs private key. The key pair is generated by the person on their own device. Content may include previously received digital speech acts being forwarded, in which case the forwarded speech act is embedded within the new one. The signature produces a unique expression verifiable by any third party using the personâs public key. The signed data includes a timestamp recording when the act was performed. These three elements combine to produce a computationally unique cryptographic stringâthe digital speech act itself. This string is a fixed expression that cannot be produced by anyone lacking the private key and can be verified by anyone with access to the public key and signed content. A digital speech act exists independently of any particular social contract or platform; it is created by a person on their own device and can be stored, transmitted, or published through any medium while retaining its cryptographic properties. The distinction between digital speech acts and superficially similar cryptographic operations becomes crucial for copyright analysis. Consider the difference between a digital speech act and a platform authentication token. A person creating a digital speech act deliberately chooses contentâ âI certify that I will deliver quality work by March 15, 2026ââand signs it, combining authentication function with expressive content. A platform authentication token, by contrast, is generated automat- ically by platform systems without user content selection: âUser ID 12345 authenticated at 14:32:07 GMT.â The authentication token serves pure function without expression; the digital speech act combines function (authentication) with deliberate expression (chosen content). This distinction matters because copyright protects expression, not mere function. C. Technical Properties The technical architecture of digital speech acts produces five characteristics relevant to their legal status and function within grassroots platforms. First, digital speech acts are independently created: each person generates their own key pair on their own device without reliance on central authority or trusted third party, maintaining autonomous control over key generation. Second, they are computationally unique: cryptographic hash functions ensure that different persons produce distinct signatures even for identical content, and the same person produces different signatures for 108 See infra Parts VII and X. 109 Ehud Shapiro, Grassroots Currencies: Foundations for Grassroots Digital Economies (2024) (arXiv preprint arXiv:2202.05619). 110 Ehud Shapiro & Nimrod Talmon, Grassroots Federation: Fair Governance of Large-Scale, Decentralized, Sovereign Digital Communities (2025) (arXiv preprint arXiv:2505.02208). 17 different content, making each signature unique with overwhelming probability to the combination of signer and content. Third, digital speech acts are permanently fixed: once created and signed, the cryptographic signature binds content to identity immutablyâany content alteration invalidates the signature and is immediately detectable. Fourth, they are verifiably attributable: attribution is cryptographically intrinsic, allowing anyone with the public key to verify the signature was produced by the holder of the corresponding private key, without dependence on platform operators, centralized databases, or intermediary trust. Fifth, they enable provenance preservation: when a digital speech act is forwarded, the original is embedded as content within the forwarderâs new digital speech act, creating cryptographic provenance chains that preserve attribution at each step and establish accountability for both original creators and subsequent forwarders. D. Two Parallel Foundations A digital speech act has two complementary characterizations: as a performative utterance in the sense of Searle and Austin, and as a volitional transaction in the sense of the volitional multia- gent atomic transactions framework. 111 These are not competing accounts but parallel foundations: speech act theory provides the linguistic-philosophical grounding establishing that the act is expres- sive; the volitional-transactions framework provides the formal grounding establishing that the act is a deliberate volition by a person, distinct from machine state. 1. Performative Utterances (Searle and Austin) Searleâs systematic taxonomy classifies illocutionary actsâutterances characterized by their illocutionary point (what the speaker intends to accomplish)âinto five categories. 112 Declarations bring about change in reality through the utterance itself (âI hereby transfer three blue coins to Sueâ; âI vote YES on Proposal #7â). Commissives commit the speaker to future action (âI promise to deliver by March 15â; âI pledge to uphold these standardsâ). Assertives commit the speaker to the truth of a proposition (âI certify this document is authenticâ; âI confirm receipt of paymentâ). Directives attempt to get the addressee to do something (âI request membership in this communityâ; âI invite Alice to join this groupâ). Expressives express the speakerâs psychological state (âI approve this proposalâ; âI thank Bob for his contributionâ). 113 Digital speech acts can instantiate any of these categories, and creating one requires substantive creative choices. The person selects the propositional content (what to say: "I vote YES on Proposal #7" versus "I vote NO"), the timing (when to perform the act), the identity (with which key pair to sign, when multiple are available), and the structural arrangement (how to format and structure the signed data). These layered dimensions of choice establish that creating a digital speech act entails expressive decisions rather than merely executing a functional operation. The personâs choice of what to say determines the illocutionary force that resultsâvoting, promising, certifyingâbut the force is a consequence of the content chosen, not a separate selection. This taxonomy builds on Austinâs foundational distinction between constative utterances, which describe states of affairs and can be evaluated as true or false, and performative utterances, which do things through being uttered rather than merely describing. 114 Digital speech acts are performative: cryptographically signing âI hereby transfer three blue coins to Sueâ does not describe a transfer but performs it. 115 The application of speech act theory to cryptographic operations is contestedâsome scholars question whether computational processes qualify as performative utterances when execution de- 111 Lewis-Pye & Shapiro, supra note 14. 112 John R. Searle, Expression and Meaning: Studies in the Theory of Speech Acts (1979). 113 For comprehensive treatment of speech act categories, see John R. Searle, A Taxonomy of Illocutionary Acts, in Expression and Meaning 1â29 (1979). 114 J.L. Austin, How to Do Things with Words (1962). 115 Cardelli et al., supra note 21. 18 pends on algorithms rather than social convention. 116 Digital speech acts differ: the person makes substantive choices about content, timing, and structure, and the cryptographic signature authenti- cates human agency. The performative force derives from deliberate human choice, not autonomous code execution. 2. Volitional Transactions (VMAT) In the volitional multiagent atomic transactions framework, each agentâs state has two compo- nents: volitionsâwhat the person wantsâand machine stateâthe state of the device. 117 A volition transaction is a free, unary act in which a single person changes their own volitions; it has no precondition. Creating a digital speech act is a volition transaction: the person decides to sign specific content, and this decision is theirs alone. The cryptographic signing is the mechanism by which the volition becomes verifiable, fixed, and attributable, but the decision precedes and grounds the mechanism. E. Digital Speech Acts as Legal Objects Having established that digital speech acts are independently created, deliberately authored, permanently fixed, and attributable to identifiable persons, the remaining question is legal rather than technical: whether existing copyright doctrine recognizes such expressions as copyrightable works. Part V addresses that question, arguing that digital speech acts satisfy the requirements of authorship, originality, and fixation under established U.S. copyright law. V. CORE LEGAL FRAMEWORK: DIGITAL SPEECH ACTS AS COPYRIGHTABLE EXPRESSION A. Introduction This Article makes the case that digital speech acts should qualify for copyright protection under existing U.S. copyright law. The argument rests on established precedent and statutory requirements. Copyright protection requires that works be original works of authorship fixed in a tangible medium of expression. 118 Digital speech acts, as cryptographically signed expressions combining content, identity, and timestamp, created through deliberate personal choices and stored on personal devices, satisfy each element. Burrow-Giles v. Sarony provides the primary doctrine for authorship: personal creative choices constitute authorship despite mechanical or algorithmic processes. Feist v. Rural establishes the minimal creativity threshold for originality that digital speech acts easily satisfy. The Copyright Actâs fixation requirement is clearly met by persistent storage on personal devices. B. What Copyright Protects in Digital Speech Acts Copyright protects expression, not ideas. 17 U.S.C. §102(b) excludes "any idea, procedure, process, system, method of operation" from protection. This foundational distinction determines what aspects of digital speech acts qualify for copyright: the ideaâauthentication, attribution, cryptographic identity verificationâis unprotectable. These are methods of operation that §102(b) deliberately excludes. The expressionâthe digital speech act as a fixed, attributable work embodying 116 See, e.g., Mireille Hildebrandt, Smart Technologies and the End(s) of Law: Novel Entanglements of Law and Technology 117â45 (2015) (questioning whether smart contracts constitute performative speech); Tiziana Terranova, Red Stack Attack! Algorithms, Capital, and the Automation of the Common, in Accelerate: The Accelerationist Reader 379â99 (Robin Mackay & Armen Avanessian eds., 2014) (analyzing performa- tivity in computational systems). 117 Lewis-Pye & Shapiro, supra note 14. 118 17 U.S.C. § 102(a). 19 specific creative choicesâis protectable. The copyrightable work is the digital speech act as a wholeâthe signed expression that results from a personâs deliberate choices. Copyright does not protect the bare cryptographic signature string considered in isolation. For given inputs (content, private key, algorithm, timestamp), the signature output is computationally determinedâthere is only one signature that can result from those specific inputs. Claiming copyright in this deterministic output would raise merger concerns: if the idea is "authenticate this specific content with this specific key" and the expression is "this unique signature string," idea and expression may merge. Copyright does protect the digital speech actâthe deliberate expressive act of signing content at a chosen time, resulting in a fixed, attributable expression. The content itself may be independently copyrightable (text, image, video), but the digital speech act adds additional creative choices: when to sign, with which identity (if multiple keys are available), what illocutionary force to convey (asserting, promising, declaring), and how to structure the signed data. These choices produce a fixed expressionâthe digital speech actâthat embodies and preserves both the content and the signing choices in verifiable form. The claim parallels Burrow-Giles: Saronyâs photograph was the product of mechanical-chemical processes following physical laws, yet copyrightable because it embodied his creative choices. 119 Similarly, the digital speech act is the product of cryptographic-mathematical processes following algorithmic rules, yet copyrightable because it embodies the personâs creative choicesâboth in content creation and in the signing act itself. The mechanical process preserves the choices; it does not negate their originality. C. Authorship under Burrow-Giles v. Sarony (1884) Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53 (1884), was the first Supreme Court case addressing whether a new technology could produce copyrightable works of authorship. Photography had existed for decades, but its mechanical nature raised fundamental questions: could photographs embody human creativity, or were they merely automatic reproductions produced by cameras rather than human authors? The Courtâs resolution established an analytical framework for evaluating whether technologies combining mechanical processes with human input produce copyrightable worksâa framework courts have applied to emerging technologies ever since. Under the Burrow-Giles framework, the determinative inquiry is not whether mechanical processes are presentâobviously they are in photographyâbut whether the work embodies "originality, of intellectual production, of thought, and conception on the part of the author." 120 The Court defined "author" constitutionally as "he to whom anything owes its origin; originator; maker." 121 Applying this framework, the Court held that photographs can be copyrighted when they embody the photographerâs "original intellectual conceptions" through creative choices. Napoleon Saronyâs photograph of Oscar Wilde was copyrightable because Sarony "made the same... entirely from his own original mental conception, to which he gave visible form by posing the said Oscar Wilde in front of the camera, selecting and arranging the costume, draperies, and other various accessories... arranging the subject so as to present graceful outlines, arranging and disposing the light and shade, suggesting and evoking the desired expression." 122 The mechanical- chemical processes of the camera preserved Saronyâs creative choices in fixed form. The presence of mechanical execution did not negate authorship; what mattered was that the resulting work embodied deliberate human creative choices. Digital speech acts present the same question Burrow-Giles answered for photography: can a new technology combining algorithmic processes with deliberate human choices produce copyrightable works? We argue the answer follows this framework. 119 Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53, 60 (1884). 120 Id. at 60. 121 Id. at 58. 122 Id. at 60. 20 Applying Burrow-Giles to digital speech acts, however, requires careful articulation because of a critical difference in how creative choices are embodied. Saronyâs creative choices were perceptibly embodied in the photograph itselfâa viewer could see the pose, lighting, and costume that Sarony selected. The creative choices were visible in the work. In contrast, a bare cryptographic signature string does not perceptibly embody timing choices or illocutionary forceâthe string does not "look like" a choice to sign now rather than later. This disanalogy, however, does not defeat copyright in digital speech acts because the copy- rightable work is not the bare signature string considered in isolation, but the digital speech act as a whole: the combination of content, cryptographic signature, and associated metadata (timestamp, public key identifier, signed data structure). This complete expression perceptibly embodies the creatorâs choices. Content embodies expressive choices. The text, image, audio, or video that the person chooses to sign contains and displays their expressive choices: what they chose to say, how they chose to say it, the illocutionary force they chose to convey. These choices are perceptible in the content itself, just as Saronyâs choices were perceptible in the photograph. The signature preserves attribution and timing. The cryptographic signature does not exist to embody choices perceptibly to human viewersâit exists to make attribution and timing crypto- graphically verifiable. When a digital speech act is displayed or transmitted, the viewer sees: (1) the content (embodying what was said), (2) the signerâs identity (embodying who said it), (3) the timestamp (embodying when it was said), and (4) the cryptographic signature that makes (2) and (3) verifiable rather than merely claimed. The signatureâs function is preservation and verification of attribution, not perceptible embodiment of aesthetic choices. Together, content plus signature plus metadata form a complete work that perceptibly embodies the creatorâs choices about what to say, when to say it, and with which identity to claim it. The signature component serves the same function in digital expression that Saronyâs signature on the photograph served: it establishes authorship and makes the attribution verifiable. The difference is that cryptographic signatures provide mathematical rather than visual verificationâbut Burrow- Giles did not require that creative choices be embodied in visually-perceptible form, only that the work embody the authorâs original intellectual conceptions. The parallel to digital speech acts is direct: just as Saronyâs creative choices constituted authorship despite the cameraâs mechanical operation, a personâs volitional creative choices constitute authorship despite cryptographyâs algorithmic operation. The constitutional definition of âauthorâ fits digital speech act creators precisely: the person possessing the private key is the exclusive originator of the signed expression; the work owes its origin to them alone. The objection that cryptographic signing is a mechanical or algorithmic process should not succeed under Burrow-Giles. Photography is mechanical; cryptography is algorithmic. Both employ technical processes that execute human creative choices. The presence of mechanical or mathematical operations does not defeat authorship when the work embodies the creatorâs original intellectual conceptionsâand digital speech acts, as performative utterances executed through cryptography, embody precisely such conceptions. D. Originality under Feist v. Rural (1991) In Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991), the Supreme Court held that Ruralâs white pages directoryâan alphabetical listing of subscriber names and phone numbersâcould not be copyrighted because it lacked the minimal creativity required for copyright protection. The Court defined originality as requiring only that the work be âindependently created by the authorâ and possess âat least some minimal degree of creativity,â emphasizing that âthe requisite level of creativity is extremely low; even a slight amount will suffice.â 123 Works need only possess âsome creative spark, âno matter how crude, humble or obviousâ it might be.â 124 Novelty is not 123 Id. at 345. 124 Id. 21 required; independent creation plus minimal creativity suffices. 125 Digital speech acts easily satisfy this minimal standard because the copyrightable workâthe digital speech act as a wholeâembodies creative choices in both its content and its creation. The person chooses what content to sign (the substance of expression), when to sign it (temporal context), which illocutionary force to employ (declaration, commissive, assertive, directive, expressive), and, if multiple identities are available, with which cryptographic identity to claim it. The resulting digital speech actâcontent plus signature plus timestampâis a fixed expression embodying these choices. The creativity resides in the choices about what is expressed, not in the cryptographic signatureâs mathematical properties. When a person creates a digital speech act stating âI vote YES on Proposal 7,â the creative choices include: deciding to vote rather than abstain, voting yes rather than no, voting on this proposal rather than another, and voting now rather than later. These choices about content and timing produce a unique expression that easily exceeds Feistâs âextremely lowâ threshold. Even the simplest digital speech act exhibits more creative choice than Ruralâs mechanical alphabetization, which Feist held to fall within âa narrow category of works in which the creative spark is utterly lacking.â 126 The signatureâs role is not to add creativity but to preserve attribution and establish verifiabil- ity. The cryptographic signature makes the creative choicesâwhat was said, when, by whomâ attributable and verifiable. Just as Saronyâs handwritten signature on his photograph established authorship without adding creative content to the image, a cryptographic signature establishes au- thorship without needing to be independently creative. The creativity lies in the deliberate expressive choices the signature preserves, not in the signatureâs mathematical derivation. Feist also clarifies the infringement standard: a plaintiff must prove âownership of a valid copyrightâ and âcopying of constituent elements of the work that are original.â 127 Because each digital speech act is cryptographically unique to the person who created it, both elements are readily established. The signature provides mathematical proof that only the holder of the private key could have created this particular signed expression. Recognizing copyright in cryptographically signed expressions raises concerns about extending protection too broadly. Mark Lemley has demonstrated how copyright protection for works with minimal creativity can lead to strategic claiming over functional elements, creating anticommons problems that restrict follow-on creation and innovation. 128 Pamela Samuelson has cautioned against courts applying minimal creativity standards so permissively that formulaic or purely functional expressions receive protection. 129 We address these concerns in Part IX, demonstrating that digital speech acts avoid the problems thin copyright creates because their person-specific cryptographic nature prevents the strategic overclaiming and functional appropriation that scholars identify as risks of minimal creativity standards. E. Fixation under 17 U.S.C. § 102(a) Digital speech acts satisfy the Copyright Actâs fixation requirement under any defensible standard. The Act requires that works be âfixed in any tangible medium of expression. . . from which they can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device.â 130 Digital speech acts are stored persistently on the creatorâs deviceâin solid-state storage, hard drives, or other non-volatile memory. They remain accessible indefinitely, can be retrieved and displayed at any time, and persist independently of RAM or other volatile storage. This clearly satisfies fixation under any reasonable interpretation of the statute. The appropriate standard for fixation comes from Cartoon Network v. CSC Holdings, which held 125 Id. at 345â46. 126 Id. at 359. 127 Id. at 361. 128 Mark A. Lemley, Property, Intellectual Property, and Free Riding, 83 Tex. L. Rev. 1031, 1059â65 (2005); Mark A. Lemley, The Economics of Improvement in Intellectual Property Law, 75 Tex. L. Rev. 989 (1997). 129 Pamela Samuelson, The Originality Standard in Copyright Law, 42 Am. J. Comp. L. 393 (1994). 130 17 U.S.C. § 102(a). 22 that copies must persist long enough to be âperceived, reproduced, or otherwise communicatedâ for âmore than transitory duration.â 131 Digital speech acts easily satisfy this standard: they are stored persistently on devices, remain accessible indefinitely until deliberately deleted, and can be perceived and communicated at any time through the device interface. Their duration far exceeds the 1.2-second buffer copies that Cartoon Network held insufficiently fixed. Even under the more permissiveâand heavily criticizedâstandard of MAI Systems v. Peak Computer, digital speech acts would satisfy fixation. MAI held that even temporary loading of software into RAM constitutes sufficient fixation. 132 This holding has been subject to substantial scholarly criticism. Pamela Samuelson argued that MAIâs reasoning was doctrinally flawed, noting that the Copyright Actâs definition of âfixedâ requires works to be âsufficiently permanent or stable to permit [them] to be perceived. . . for a period of more than transitory duration,â a requirement that RAM copyingâwhich is lost when a computer powers offâarguably fails to meet. 133 Jessica Litman similarly criticized MAI for creating copyright liability in routine computer operations that Congress did not intend to regulate. 134 We do not rely on MAIâs permissive standard. Digital speech acts satisfy the more demanding Cartoon Network standard because they persist in non-volatile storage for indefinite duration, not merely in temporary RAM. If courts were to apply MAIâs approach, digital speech acts would satisfy fixation a fortioriâbut the better view, supported by both statutory text and scholarly criticism, recognizes that true fixation requires meaningful persistence, which device-based storage of digital speech acts provides. Whether courts apply Cartoon Networkâs âmore than transitory durationâ standard, MAIâs per- missive RAM-copying approach, or adopt an even more restrictive fixation doctrine, digital speech acts satisfy the requirement. They are embodied in tangible media (device storage), persist for more than transitory duration (indefinitely until deleted), and can be perceived and communicated at any time through the device interface. F. Synthesis: Copyright and Architecture This Article has argued that digital speech acts satisfy the traditional requirements of copyrighta- bility. Under Burrow-Giles, they embody personal creative choices despite the use of algorithmic processes. Under Feist, those choices easily satisfy the minimal originality threshold. And because digital speech acts are persistently stored on personal devices, they satisfy the Copyright Actâs fixation requirement. If that analysis is accepted, an important implication follows. The same body of copyright law that helped enable platform dominance also contains doctrines that support distributed ownership. Part I showed how centralized platforms rely on copyright protection for software, expansive licensing arrangements, and adhesion contracts to aggregate control over user activity. Digital speech acts invoke a different strand of copyright doctrine: authorship, originality, and fixation. The difference is not a difference in law but a difference in architecture. Platform systems are organized around centralized servers controlled by intermediaries. Digital speech acts are created and stored on individualsâ own devices. Those architectural choices determine which copyright doctrines become practically significant and who is positioned to exercise the rights copyright confers. The significance of digital speech acts, therefore, is not merely that they may be copyrightable works. It is that they provide a mechanism through which copyright ownership can remain with the individual creator rather than being absorbed into platform-controlled systems. Copyright law alone cannot guarantee that result. Rights may exist formally while becoming ineffective in practice through architectural arrangements that separate ownership from possession and control. 131 Cartoon Network LP v. CSC Holdings, Inc., 536 F.3d 121, 127â30 (2d Cir. 2008). 132 MAI Systems Corp. v. Peak Computer, Inc., 991 F.2d 511, 518 (9th Cir. 1993). 133 Pamela Samuelson, The Copyright Grab, Wired (Jan. 1996), https://archive.gyford.com/1997/wired- uk/2.01/features/copyright.html; see also Jessica Litman, The Exclusive Right to Read, 13 Cardozo Arts & Ent. L.J. 29, 34â37 (1994) (criticizing MAIâs expansive fixation holding as inconsistent with the Copy- right Actâs text and purpose). 134 Id. 23 Recognizing digital speech acts as copyrightable expression establishes a legal foundation for distributed ownership. If digital speech acts qualify as original works of authorship, then the individual who creates them occupies the position of author within the copyright system and possesses the exclusive rights that copyright law attaches to authorship. The question that follows is therefore not whether digital speech acts can be owned, but who the law recognizes as the relevant rights- bearing entity. Contemporary digital systems routinely collapse individuals into platform-defined categories such as âusers,â treating participation as a function of access to services rather than as an exercise of independent legal agency. Digital speech acts suggest a different framing. Copyright law attaches authorship to the person who originates expression, not to the intermediary through which expression is communicated. The next Part examines this distinction between the person and the user and explains why it is foundational to digital sovereignty. VI. THE USER/PERSON DISTINCTION This Article consistently distinguishes between users and persons. The distinction is not stylistic but analytical, reflecting how different architectural arrangements construct different legal subjects and enable different legal relationships. We use the term âusersââwith all the legal and economic implications that term car- riesâwhen discussing centralized platforms. Platform participants are users: a discrete legal category created by software licensing, defined by license-bound access, adhesive agreements, asymmetric obligations, and possession-surrender. We use the term âpersonsââboth natural persons (human beings) and juridical persons (corporations, organizations)âwhen discussing grassroots networks. Grassroots participants are persons: owners of devices, possessors of private keys, creators of copyrighted expressions, and parties to voluntary digital social contracts. The distinction tracks a fundamental architectural difference that produces different legal subjects. A. âUserâ as a Legal Category Created by Platform Architecture A user is a discrete legal category that emerges from software licensing applied at platform scale. The term entered legal discourse through the software industryâs characterization of people who access and use software as distinct from those who own it. 135 This distinction was crucial to circumventing copyrightâs first-sale doctrine: if transactions are characterized as licenses rather than sales, recipients are "users" bound by license terms rather than "owners" entitled to copyright exemptions. 136 The "user" as a legal subject is defined by four characteristics, each carrying specific legal and economic implications: (1) License-bound status: Users access software and platforms only through revocable licenses, never as owners. Legal implication: Users possess no first-sale rights, no right to transfer access, and no protection against arbitrary termination. Economic implication: Platforms can extract ongoing rent for access that would otherwise require only a one-time purchase. 137 (2) Adhesive agreement: Users must accept non-negotiable terms drafted unilaterally by software providers. Legal implication: No meaningful consentâusers cannot negotiate, modify, or 135 See Madison, supra note 54, at 280â85 (tracing emergence of âuserâ as distinct from âownerâ through software licensing practice). 136 See 17 U.S.C. § 109(a) (first sale doctrine applies to "owner of a particular copy"); MAI Systems Corp. v. Peak Computer, Inc., 991 F.2d 511, 518 n.5 (9th Cir. 1993) (software recipients are "licensees" not "owners" and thus not entitled to § 117 exemptions). 137 See ProCD, Inc. v. Zeidenberg, 86 F.3d 1447, 1450â52 (7th Cir. 1996) (enforcing license prohibiting commercial use despite purchaser buying retail copy). 24 reject specific terms while maintaining access. Economic implication: Platforms can impose any terms markets will bear, including comprehensive content licenses and liability waivers. 138 (3) Asymmetric obligations: Users grant broad rights (content licenses, data access, behavioral tracking) while receiving only revocable access. Legal implication: Platforms obtain transfer- able, sublicensable, royalty-free licenses to user content while users receive non-transferable access subject to immediate termination. Economic implication: Platforms monetize user- generated content (advertising, training data, derivatives) without compensation to creators. 139 (4) Surrender of possession: Users upload content to provider-controlled servers, relinquishing physical possession while retaining nominal copyright ownership. Legal implication: Plat- forms control access, modification, distribution, and preservation of user content regardless of copyright ownership. Economic implication: Users cannot extract value from their own content without platform permission and mediation. 140 Platforms operationalize this legal category at global scale. Metaâs 3.58 billion Family Daily Active People (DAP) as of December 2025, 141 YouTubeâs more than 2 billion monthly logged-in users, 142 and Xâs hundreds of millions of users all occupy the "user" subject position: license-bound, adhesively-contracted, asymmetrically-obligated, possession-surrendering participants in platform- controlled infrastructure. The genealogy reveals that the modern âuserâ is not a natural legal category but a product of software licensing. 143 Prior to the rise of software licensing, purchasers of books, records, and other copyrighted works were generally treated as owners rather than users. The first-sale doctrine entitled them to resell, lend, or dispose of their copies without copyright-holder permission. 144 Software licensing disrupted this arrangement by distinguishing licensees from owners. 145 Courts largely accepted this distinction, 146 and platforms later generalized it to billions of participants in digital environments. B. Personhood in Grassroots Architectures In contrast, grassroots social networking has no architectural position for an entity to construct "users" as a legal category. With no platform to draft Terms of Service, no servers to which content must be uploaded, and no central authority conditioning access on license grants, the relevant legal subject is the personânatural or juridicalâwho participates through devices they own and expressions they create. In U.S. law, "person" includes both natural persons (human beings) and juridical persons (cor- porations, companies, associations, partnerships, and other legal entities). 147 Both natural and juridical persons can participate in grassroots networks. A corporation could operate through devices it owns, authenticate through private keys it possesses, and create digital speech acts it copy- rightsâjust as a human being could. The legal category is "person" not "user" because no platform exists to construct the user-subject through license-mediation. A person in grassroots architecture is defined by four characteristics that mirror and invert the user-subject position: 138 See Nancy S. Kim, Wrap Contracts: Foundations and Ramifications 47â73 (2013) (analyzing adhesive nature of software licenses). 139 See infra Part I.E (examining contemporary platform Terms of Service). 140 See Cohen, supra note 12, at 37â65 (platforms obtain control through possession regardless of nominal ownership). 141 Meta 2025 Form 10-K, supra note 88 142 YouTube, Press Page, https://w.youtube.com/about/press/ (last visited May 15, 2026). 143 For a detailed genealogy of this development, see infra note 10. 144 17 U.S.C. § 109(a); Bobbs-Merrill Co. v. Straus, 210 U.S. 339, 350 (1908). 145 See Madison, supra note 54, at 287â94. 146 MAI Systems, 991 F.2d at 518 n.5; ProCD, 86 F.3d at 1450â52. 147 1 U.S.C. § 1 ("the word âpersonâ ... include[s] corporations, companies, associations, firms, partnerships, societies, and joint stock companies, as well as individuals"). 25 (1) Ownership not license: Persons own the devices on which they operate (smartphones for natural persons, servers for juridical persons), the private keys through which they authenticate, and the copyright in expressions they create. Legal implication: Full first-sale rights, transferability, and protection against arbitrary deprivation. Economic implication: Persons can extract value from their own property and expressions without intermediary permission or rent extraction. (2) Voluntary agreement not adhesion: Persons participate through digital social contractsâ voluntary agreements among participants, specified and enforced through code. 148 Legal implication: Genuine consentâparticipation is optional, terms are visible in code, and exit carries no penalty beyond loss of access to that particular network. Economic implication: No entity can impose extractive terms as a condition of participation. (3) Symmetric obligations: Persons entering digital social contracts accept the same rules as all other participants. Legal implication: No legal asymmetry between entity (which extracts) and participant (who surrenders). All participants have equal legal standing. Economic implication: No entity positioned to extract value from othersâ contributions without providing equivalent value. (4) Retained possession: Persons store their own data on their own devices. Nothing is uploaded to third-party servers; nothing is relinquished to othersâ possession. Legal implication: Copyright ownership and physical possession coalesceâthe person who owns the copyright also possesses the copyrighted work. Economic implication: No intermediary can monetize, manipulate, or control access to the personâs content. C. Why the Distinction Matters The distinction between users and persons is not merely terminological. It identifies two different legal subject positions produced by two different architectures. Much contemporary scholarship seeks to improve the position of users within platform systems through stronger privacy protections, portability rights, or limitations on contractual overreach. Such proposals generally accept the continued existence of platform-mediated relationships and focus on modifying the terms under which users participate. The present analysis instead asks what legal consequences follow when the architectural conditions that produce the user-subject are absent altogether. Grassroots architectures do not merely offer users better terms. They eliminate the architectural conditions through which users are constructed as a distinct legal category. Where platforms position individuals as licensees operating through accounts, grassroots systems position participants as persons operating through devices they own, keys they control, and expressions they author. The distinction becomes particularly significant when copyright ownership is considered. Copy- right law vests rights in authors and owners, not in users as such. If digital speech acts qualify as copyrightable expression, the question is not whether a platform user possesses rights granted by a service provider, but whether a person possesses rights arising directly from authorship. The answer determines where ownership resides, who exercises control over expression, and whether copyright functions as a mechanism of dependence or autonomy. The distinction between user and person therefore serves as the conceptual bridge between architecture and ownership. Copyright ownership thus becomes more than a private economic entitlement: it becomes a legal mechanism through which persons may exercise control over their digital expressions and the means through which those expressions are attributed, authenticated, and shared. VII. COPYRIGHT OWNERSHIP AND DIGITAL SOVEREIGNTY Part I introduced digital sovereignty as the ability to conduct oneâs digital life free of third- party control, surveillance, manipulation, and rent-seeking. Part VII demonstrates why copyright 148 Cardelli et al., supra note 21. 26 ownership is a prerequisite for achieving that sovereignty in grassroots architectures. Copyright establishes legal ownership that complements the cryptographic possession inherent in device-based systems, closing the possession-ownership gap that platforms exploit through licensing extraction. The claim is not merely that copyright protection for digital speech acts would be beneficial. The claim is structural: grassroots architectures can provide cryptographic security and technical control, but without copyright recognition, people cannot enforce ownership claims through legal institutions or prevent unauthorized commercial exploitation. Copyright transforms technical possession into legally enforceable ownership, enabling digital sovereignty. A. Dual Ownership: Cryptographic Possession and Legal Copyright Digital speech acts implicate two complementary forms of authority, neither alone sufficient but together constitutive of digital sovereignty: cryptographic possession and legal ownership. Cryptographic possession. When a person creates a digital speech act, they exercise factual control through private key cryptography. Only the person possessing the private key can produce valid signatures. This creates a form of technical possession enforced by mathematical impossibility of forgeryâno intermediary can create, modify, or repudiate a digital speech act without the private key. The expression is stored on the personâs own device, under their exclusive physical control. Cryptographic possession is factual control over a digital object, not legal possession in property lawâs doctrinal sense. 149 A person may have exclusive technical control over data on their device while lacking legal rights to prevent others from copying, distributing, or commercially exploiting that data if they can obtain it through other means. Cryptographic possession provides control and security, but not legal ownership. Legal copyright. Copyright provides legal ownership but, standing alone in platform architec- tures, proves insufficient for digital sovereignty. A person may retain copyright in content they create while being structurally compelled to license that copyright to platforms as the condition of partici- pation. Legal ownership without control over infrastructure creates the dependency relationship that enables platform extraction. If digital speech acts qualify as copyrightable works under Part Vâs analysis, that legal owner- ship takes the form of the exclusive rights enumerated in Section 106: reproduction, preparation of derivative works, distribution, public performance, and public display. 150 These rights are enforce- able through legal institutions against third parties who copy, distribute, or commercially exploit the work without authorization. Dual ownership in grassroots architectures. When people create and store digital speech acts on their own devices using their own private keys, and copyright law recognizes them as authors of those expressions, cryptographic possession and legal ownership converge. The person has both factual control (technical possession via private key) and legal rights (copyright ownership via authorship). No intermediary controls the infrastructure; no intermediary can compel licensing; no license must be granted as condition of participation. This dual ownershipâtechnical and legalâis what distinguishes grassroots from platform archi- tectures. Grassroots architectures unify possession and ownership: the person possesses the content (stored on personal device) and owns it legally (copyright via authorship), with no intermediary posi- tioned to extract licensing rights. The convergence of cryptographic possession and legal ownership closes the possession-ownership gap identified earlier and provides the ownership structure on which digital sovereignty depends. B. From Ownership to Sovereignty Dual ownership establishes the conditions for digital sovereignty because it prevents the concen- tration of control that characterizes platform architectures. Where possession and ownership converge 149 On the distinction between factual control and legal possession, see Carol Rose, Possession as the Origin of Property, 52 U. Chi. L. Rev. 73 (1985); Henry E. Smith, Property as the Law of Things, 125 Harv. L. Rev. 1691 (2012). 150 17 U.S.C. § 106. 27 in the same person, no intermediary occupies the position necessary to control participation, monitor activity, extract rents, or dictate the terms of digital life. Device-based architecture provides the technical foundation: individuals control the infrastruc- ture through which they communicate and authenticate. Copyright provides the legal foundation: individuals own the expressions they create and may enforce that ownership against unauthorized appropriation. Together, these forms of control align possession and ownership in the same person. This alignment matters because the defining feature of contemporary platform architectures is the concentration of control. Platforms aggregate control over infrastructure, communication channels, data storage, and the practical use of copyrighted expression. Even where users formally retain copyright ownership, platform control over possession and access allows platforms to monitor activity, dictate terms of participation, extract licensing rights, and mediate social relationships. The result is a structural dependency in which digital participation occurs only through infrastructure controlled by others. Grassroots architectures alter that allocation of control. Control is distributed across participants rather than concentrated in a platform. Copyright ownership does not eliminate all forms of interfer- ence, nor does cryptographic possession eliminate every practical vulnerability. What they eliminate is the structural dependency created when communication depends upon infrastructure owned and governed by third parties. Digital sovereignty therefore follows not from ownership alone but from the distribution of ownership. A person who both possesses and owns their digital expressions is no longer situated as a licensee operating through anotherâs infrastructure. They participate as an owner operating through infrastructure they control. The significance of copyright recognition is thus not merely economic. Copyright transforms digital speech acts from expressions that can be technically controlled into expressions that can also be legally owned. Combined with device-based architecture, that ownership structure disperses control across participants and removes the centralized points of authority through which surveillance, manipulation, and rent-seeking become possible. If digital speech acts do not qualify as copyrightable works, grassroots architectures may pro- vide cryptographic security but not legal ownership. Participants would possess their expressions technically without possessing enforceable rights against appropriation or exploitation. Copyright recognition completes the ownership structure. By aligning legal ownership with cryptographic possession and distributing both across participants, it provides a necessary foundation for digital sovereignty. The democratic and economic implications of this distributed ownership structureâincluding its relationship to political autonomy, democratic participation, and the governance of digital spacesâare explored later in Part X. VIII. PROVENANCE, ACCOUNTABILITY, AND LEGAL CULPABILITY Part V established that digital speech acts should qualify as copyrightable worksâspecifically, that the combination of content, cryptographic signature, and metadata constitutes a copyrightable expression. Ownership, however, does not resolve questions of responsibility. A system in which individuals own their expressions but cannot determine who created, modified, or redistributed them remains vulnerable to manipulation, misinformation, and attribution disputes. Digital sovereignty requires not only distributed ownership but also distributed accountability. The same cryptographic architecture that aligns possession and ownership also creates verifiable provenance. Every digital speech act is signed by its creator, and every forwarding event creates a new signed expression that incorporates the prior one. Provenance therefore records not only authorship but also the history of dissemination. It becomes possible to identify who created content, who redistributed it, and in what sequence. This Part examines the legal consequences of that structure. What happens when digital speech acts are forwarded through chains of participants, each adding a new cryptographic signature? Does the original creatorâs copyright persist through successive forwarding? What legal responsibility do forwarders bear for content they redistribute? How are provenance chains authenticated as evidence? 28 And how do questions of jurisdiction arise when forwarding crosses national borders? A. Cryptographic Provenance Chains A digital speech act is either an original expressionâa person deliberately signs content with their private keyâor a forwarding expressionâa person signs a new digital speech act whose content consists of a previously received digital speech act. In the forwarding case, the forwarded speech act is embedded within the new one, including the original creatorâs signature. When content is forwarded multiple times, each forwarding creates a new digital speech act that embeds the preceding one. The result is a chain: the outermost speech act is signed by the most recent forwarder and contains, as content, the speech act of the previous forwarder, which in turn contains the speech act before it, continuing back to the original creatorâs speech act. Each link in the chain is cryptographically signed by the person who created it, and the chain as a whole constitutes a verifiable record of the contentâs provenanceâwho created it, who forwarded it, and in what order. 151 Unlike conventional platform architectures, where dissemination histories are typically main- tained by platform operators and may be incomplete, inaccessible, or subject to alteration, prove- nance in a digital speech act is embedded directly within the expression itself. The record of creation and dissemination therefore travels with the content wherever it goes, creating a persistent and independently verifiable history of the expressionâs origin and transmission. B. Copyright Preserved at Each Step From a copyright perspective, forwarding is additive rather than substitutive. The original digital speech act remains fixed and intact within the forwarding speech act, together with the cryptographic signature that identifies its author. The forwarder creates a new digital speech act incorporating the earlier one, but does not displace or alter the original work. As a result, the chain contains multiple copyrightable expressions layered together, each retaining its own authorship and legal status within the provenance chain. This structure is analogous to republishing a book with a new foreword. The foreword may itself be copyrightable, but its creation does not diminish or replace the copyright in the underlying book. Likewise, each forwarding digital speech act may constitute a new copyrightable expression, while the embedded digital speech acts retain their independent copyrights. The cryptographic structure makes this relationship explicit and verifiable: the original creatorâs signed expression remains embedded within every subsequent forwarding, regardless of how many times the content is redistributed. C. Accountability and Legal Liability Each digital speech act is cryptographically signed by its creator, establishing verifiable authorship and deliberate choice to create or forward content. This cryptographic accountability has complex legal implications that differ substantially from platform architecturesâin some ways increasing individual responsibility, in others raising unresolved questions about the application of existing intermediary liability frameworks. 1. Forwarder Liability: Unresolved Questions Grassroots architectures eliminate platform intermediaries, raising the question: what liability do individual forwarders face? The answer is doctrinally complex and depends on how courts apply existing frameworks to peer-to-peer forwarding. Section 230 and grassroots participants. Section 230âs immunity extends to âusersâ of inter- active computer services, not only to service providers. 152 Courts have applied this protection to 151 See Shapiro, supra note 15 (âutterances that have been forwarded multiple times are provided with their entire provenanceâ). 152 47 U.S.C. § 230(c)(1) (âNo provider or user of an interactive computer service....â) (emphasis added). 29 individuals who republish content created by others online. 153 In Barrett, the California Supreme Court held that an individual who redistributed allegedly defamatory messages qualified as a pro- tected âuserâ under Section 230. 154 Whether a person participating in a grassroots provenance network would be treated as a "user of an interactive computer service" for purposes of Section 230 remains unresolved. Section 230âs text extends protection to "users" of interactive computer services, and courts have applied that protection to individuals who republish third-party content online. On one view, a person who forwards a digital speech act may fall within that category in much the same way that the defendant in Barrett was treated as a user of Usenet. On another view, grassroots architectures differ in legally significant respects. Because grassroots participants act simultaneously as creators, distributors, and recipients rather than through a centralized intermediary, existing Section 230 doctrine provides only an imperfect analogue. Section 230 does not protect actors who materially contribute to allegedly unlawful content or who become "information content providers" through the creation or development of the challenged expression. 155 A participant who adds commentary, selectively redistributes content to amplify harmful material, or knowingly propagates falsehoods may fall outside the statuteâs protection. The cryptographic provenance structure introduces an additional uncertainty. Every forwarding event is affirmatively signed, creating a verifiable record that a particular person chose to redistribute the content. Courts may conclude that such deliberate, attributable forwarding remains protected republication activity under Section 230. Alternatively, they may view cryptographically authenti- cated forwarding as evidence of a level of editorial participation sufficient to place the actor outside the statuteâs intended protections. Existing doctrine provides no definitive answer. Accordingly, the application of Section 230 to grassroots forwarding remains unresolved. The provenance chain strengthens attribution and accountability, but whether those same features preserve or diminish statutory immunity is ultimately a question for future courts. DMCA safe harbors and grassroots participation. The DMCAâs Section 512 safe harbors protect certain âservice providersâ from copyright infringement liability arising from the activities of others, provided they satisfy specified statutory requirements. 156 The statute defines a service provider broadly to include âan entity offering the transmission, routing, or providing of connections for digital online communications.â 157 Whether participants in a grassroots architecture fall within this definition is uncertain. Courts have applied Section 512 to a variety of intermediaries, but the paradigm cases involve entities operating communications infrastructure or online services for the benefit of others. 158 A person who receives and forwards digital speech acts through their own device occupies a different position. Such a participant is not obviously providing a communications service to the public; rather, they are engaging in expressive activity through the creation and forwarding of digital speech acts. Accordingly, it is unclear whether Section 512âs safe harbors would apply to grassroots par- ticipants. Courts may conclude that forwarding activity falls outside the category of protected intermediary conduct, leaving questions of contributory or vicarious copyright infringement to be resolved under traditional copyright doctrines. Defamation and the republication rule. The traditional tort rule holds that â[o]ne who repeats or otherwise republishes defamatory matter is subject to liability as if he had originally published it.â 159 153 See Batzel v. Smith, 333 F.3d 1018, 1030â32 (9th Cir. 2003) (operator of email listserv qualified for § 230 immunity); Barrett v. Rosenthal, 40 Cal. 4th 33, 41â61 (2006) (individual who reposted allegedly defamatory messages to Usenet newsgroup entitled to § 230 immunity). 154 Barrett, 40 Cal. 4th at 51â54. 155 See Fair Hous. Council of San Fernando Valley v. Roommates.com, LLC, 521 F.3d 1157, 1168â69 (9th Cir. 2008) (en banc) (no immunity when website materially contributed to alleged illegality). 156 17 U.S.C. § 512. 157 17 U.S.C. § 512(k)(1)(B). 158 See, e.g., Ellison v. Robertson, 357 F.3d 1072, 1076â80 (9th Cir. 2004) (applying § 512 to Internet access provider). 159 Restatement (Second) of Torts § 577(1) (1977). 30 This suggests that a person who forwards a defamatory digital speech act may incur liability through the act of republication itself. However, the single-publication rule complicates that conclusion. The single-publication rule provides that âthe publication of a writing... gives rise to only one claim for relief for libel.â 160 The rule was developed to prevent mass-media defendants from facing separate suits for every copy of a newspaper or book distributed. 161 Courts later extended the doctrine to online publications, treating the posting of content on a website as a single publication regardless of the number of subsequent views. 162 Whether the forwarding of a digital speech act constitutes a new publication or merely extends an existing one is uncertain. On one view, forwarding resembles the continued circulation of an already published work and therefore falls within the logic of the single-publication rule. On another view, each forwarding event intentionally transmits the expression to a new audience and therefore constitutes a distinct republication. Courts have long distinguished between passive availability and active redistribution. 163 Deliberate peer-to-peer forwarding appears more analogous to the latter, but existing doctrine provides no definitive answer. Existing defamation law was developed in the context of publishers, broadcasters, and platform- hosted communications rather than cryptographically attributable provenance chains. As a result, courts could conclude that forwarding creates independent republication liability, or they could treat forwarding as part of a broader continuing publication. The provenance chain does not determine the legal outcome, but it does provide unusually clear evidence regarding the conduct at issue. Each forwarding event is independently attributable to a particular person and can be verified cryptographically. The legal significance of that attribution remains uncertain, but the evidentiary record is substantially clearer than in conventional online environments. Copyright and contributory infringement. A person who forwards infringing content may face contributory copyright infringement liability. The traditional test requires: (1) knowledge of the infringing activity, and (2) material contribution to the infringement. 164 A participant who knowingly forwards infringing content contributes to its continued distribution, potentially satisfying both elements depending on the circumstances of the forwarding and the extent of the contribution. Separate from contributory infringement is inducement liability. In MGM Studios Inc. v. Grokster, Ltd., the Supreme Court held that âone who distributes a device with the object of promoting its use to infringe copyright... is liable for the resulting acts of infringement by third parties.â 165 Inducement requires proof of purposeful, culpable expression and conduct directed toward encouraging infringement. 166 Participants who merely forward content are unlikely to incur inducement liability absent evidence that they encouraged or promoted infringement, but the doctrine remains available in more egregious cases. The provenance chain does not determine whether liability exists, but it does provide evidence relevant to the inquiry. Because each forwarding event is independently signed, the chain identifies the sequence of participants through which content was distributed. This may assist in establishing questions of knowledge, participation, and causation that often prove difficult in conventional online environments. At the same time, liability remains dependent on the specific facts of each case. A person who forwards content without knowledge of infringement may lack the requisite scienter, while a person whose forwarding plays only a minimal role in dissemination may not satisfy traditional standards of material contribution. 160 Firth v. State, 98 N.Y.2d 365, 371 (2002); see also Restatement (Second) of Torts § 577A (1977). 161 See Keeton v. Hustler Magazine, Inc., 465 U.S. 770, 777 n.8 (1984) (describing the single-publication rule). 162 See Firth, 98 N.Y.2d at 371â72 (single-publication rule applies to online postings). 163 See Hoffman v. Capital Cities/ABC, Inc., 255 F.3d 1180, 1185 (9th Cir. 2001) (passive continued availability of online content does not create new publication; republication may occur when material is directed to a new audience). 164 Gershwin Publâg Corp. v. Columbia Artists Mgmt., Inc., 443 F.2d 1159, 1162 (2d Cir. 1971). 165 545 U.S. 913, 919 (2005). 166 Id. at 936â37. 31 2. Accountability and Chilling Effects The foregoing discussion reveals a tension inherent in attributable communication systems. The same provenance mechanisms that provide transparency and accountability may also increase perceived legal risk. If participants believe that forwarding a digital speech act could expose them to defamation, copyright, or other claims, some may choose not to participate in forwarding at all. Existing defenses remain available, including truth in defamation actions, fair use in copyright disputes, and the absence of knowledge where liability doctrines require scienter. Moreover, prove- nance chains may assist not only plaintiffs but also defendants by providing evidence regarding the origin of content, the scope of dissemination, and the participantâs role within the chain. Nonetheless, uncertainty itself can produce caution. Existing liability doctrines developed in environments characterized by publishers, intermediaries, and platform-hosted communications. Their application to cryptographically attributable provenance chains remains unsettled. Until courts clarify how these doctrines apply, some participants may perceive forwarding as carrying greater legal risk than comparable conduct on platform-mediated networks. Whether this increased accountability ultimately discourages participation or instead promotes more deliberate and responsible dissemination practices is an empirical question beyond the scope of this Article. The important point for present purposes is that provenance changes not only the evidentiary landscape but also the incentives surrounding redistribution. 3. Accountability Through Provenance The preceding discussion does not yield definitive answers regarding liability. Existing doctrines governing intermediary immunity, republication, and contributory infringement developed in legal environments that did not contemplate cryptographically attributable provenance chains. Whether courts extend, modify, or limit those doctrines in the context of digital speech acts remains uncertain. What is less uncertain is the effect of provenance on accountability. Provenance chains provide a transparent record of authorship and dissemination that is largely absent from contemporary platform-mediated communications. This transparency does not itself determine liability. A provenance chain cannot establish whether a statement is defamatory, whether a participant possessed the requisite knowledge for contributory infringement, or whether a particular immunity applies. Those questions remain matters of substantive law. Provenance instead affects the evidentiary foundation upon which those questions are resolved. In this respect, grassroots architectures invert a defining characteristic of contemporary platforms. Platform systems frequently rely upon centralized moderation, opaque recommendation mechanisms, and institutional records to determine responsibility. Provenance-based architectures distribute expressive authority among participants while simultaneously preserving a verifiable record of their actions. The result is not immunity from legal responsibility but a framework in which responsibility can be traced with greater precision. Whether this shift ultimately advances democratic values, free expression, or effective governance is a broader normative question addressed in subsequent Parts. For present purposes, the relevant point is narrower: digital speech acts transform accountability from an inference drawn from platform records into a property embedded within the architecture itself. D. Evidentiary Authentication of Provenance Chains The accountability value of provenance depends upon more than technical verifiability. Prove- nance chains may record authorship, forwarding, and dissemination with cryptographic precision, but those records must still be admissible in legal proceedings before they can support legal rights or obligations. The question therefore becomes how a provenance chain proves its authenticity to a court. This issue is central to any litigation over digital speech acts and requires examination of authentication requirements under the Federal Rules of Evidence. 32 1. The Authentication Requirement To be admissible, evidence must be authenticated: the proponent must âproduce evidence suffi- cient to support a finding that the item is what the proponent claims it is.â 167 For a digital speech act, authentication requires establishing that: (1) the content was created or forwarded by the person whose signature appears on it, and (2) the content has not been altered since signing. Traditional authentication of digital evidence involves witness testimony (âI sent this emailâ), distinctive characteristics (âthis appears to be from the defendantâs email accountâ), or chain of custody evidence. 168 Digital speech acts present a different model of authentication: cryptographic proof that a signature is mathematically valid for a given public key. 2. Self-Authenticating Digital Signatures Rule 902(14), added in 2017, permits self-authentication of electronically stored information authenticated through a reliable process of digital identification. The Advisory Committee Notes specifically contemplate authentication through hash values and similar cryptographic techniques. A digital speech actâs cryptographic signature operates similarly: it is a mathematical func- tion of the content and private key that can be verified using the corresponding public key. If verification succeeds, the signature mathematically proves the content has not been altered since signing and was signed by whoever holds the private key corresponding to the public key. This processâcryptographic verificationâis a âprocess of digital identificationâ under Rule 902(14). 3. Authentication Procedure Authentication under Rule 902(14) would require several related showings. First, the proponent would present the digital speech act itself, including the content, signature, timestamp, and public-key identifier. Second, the proponent would demonstrate that cryptographic verification of the content, signature, and public key produces a valid result, establishing that the signed content has not been altered and that it was signed using the corresponding private key. Third, the proponent would need to link the public key to the person alleged to have created or forwarded the digital speech act. This remains the critical evidentiary step. Cryptographic verification can establish that a particular key signed a particular message, but additional evidence is ordinarily required to establish who controlled that key. Such evidence may include testimony, prior use of the public key in communications, public registration of the key, admissions by the purported signer, or circumstantial evidence connecting the content to a particular individual. Finally, Rule 902(14) contemplates certification of the digital-identification process and its re- liability. Once these elements are established, the digital speech act may be authenticated through cryptographic verification without requiring the traditional forms of extrinsic authentication com- monly associated with electronic evidence. 4. Provenance Chain as Business Record Alternatively, a provenance chain may be admissible through the business-records exception under Rule 803(6) where records of digital speech acts are systematically maintained in the regular course of an organizationâs activities. 169 The cryptographic signatures and timestamps would be records âkept in the course of a regularly conducted activityâ if the grassroots application automatically records them. This approach requires testimony from a custodian or qualified witness that the records are maintained systematically and reliably. However, the business record approach faces limitations: the hearsay exception addresses the record-keeping process, not the truth of the content. A business record establishes that a signature was recorded, not that the signature is valid. Cryptographic authentication under Rule 902(14) is stronger: it proves mathematical validity of the signature, not merely that someone recorded it. 167 Fed. R. Evid. 901(a). 168 See Fed. R. Evid. 901(b)(1), (4), (9). 169 Fed. R. Evid. 803(6). 33 5. Challenges and Limitations Cryptographic authentication significantly simplifies evidentiary questions, but it does not elimi- nate them. A valid digital signature proves that a particular private key signed a particular expression and that the signed content has not been altered. It does not, by itself, resolve every factual question relevant to litigation. Linking public keys to persons. The most significant limitation is attribution. Signature verification establishes that a digital speech act was signed by the private key corresponding to a particular public key, but a court must still determine who controlled that key. If a defendant denies ownership, the proponent must establish the connection through testimony, prior communications, public registration records, admissions, or other circumstantial evidence. This challenge resembles traditional authentication disputes over email accounts or social media profiles: proving that a communication originated from a particular account does not automatically prove who controlled the account at the relevant time. 170 Key compromise. A defendant may contend that a private key was lost, stolen, or otherwise compromised. Cryptographic verification cannot independently disprove such a claim; it establishes only that the key was used. As with traditional allegations of forgery or unauthorized account access, the factfinder must evaluate the credibility of the claim in light of surrounding evidence, including the circumstances of the communication and the defendantâs prior use of the key. Timestamp reliability. Digital speech acts may include timestamps, but the signature proves only that the timestamp formed part of the signed expression. It does not necessarily establish that the recorded time is accurate. If the device clock was incorrect or manipulated, additional evidence may be necessary to verify when the communication actually occurred. Corroborating evidence might include receipt by other participants, contemporaneous records, or references within the content itself. Expert testimony and technical understanding. Although Rules 901 and 902 permit au- thentication through digital identification processes, courts and juries may still require assistance understanding how cryptographic verification operates. Expert testimony may therefore be neces- sary to explain the underlying technology, the significance of a valid signature, and the limits of the conclusions that can properly be drawn from it. 171 These limitations define the scope of what cryptographic authentication establishes. A prove- nance chain can provide unusually strong proof that a particular expression was signed by a particular key and remained unchanged thereafter. Questions of key ownership, authorization, and surrounding context remain matters for conventional evidentiary proof. 6. Comparative Advantage Over Platform Evidence Despite these limitations, provenance chains offer significant evidentiary advantages over con- ventional platform-hosted communications. Authentication of platform content typically depends upon records maintained by the platform itself. A party seeking to authenticate a social-media post, message, or account activity often relies on business records, account logs, metadata, or testimony supplied by the platform operator. Although such evidence may be sufficient under traditional evidentiary standards, its reliability ultimately depends upon the accuracy of the platformâs record-keeping systems and the availability of the platform to provide the relevant information. Digital speech acts shift much of this authentication burden from institutional record keeping to cryptographic verification. A valid digital signature directly demonstrates that a particular public key signed a particular expression and that the signed content has not been altered thereafter. Authentication therefore depends less upon the records of a third-party intermediary and more upon mathematical verification of the communication itself. 170 See, e.g., United States v. Siddiqui, 235 F.3d 1318, 1322â23 (11th Cir. 2000) (authentication of email requires proof of authorship). 171 See, e.g., In re Vee Vinhnee, 336 B.R. 437, 444â45 (9th Cir. BAP 2005); United States v. Simpson, 152 F.3d 1241, 1250 (10th Cir. 1998). 34 This distinction does not eliminate ordinary evidentiary disputes. Questions of key ownership, authorization, compromise, and attribution remain subject to proof through conventional means. But once the public key is linked to a person, the provenance chain provides unusually strong evidence of integrity and transmission history without requiring reliance upon a platformâs internal records. The evidentiary advantages become particularly significant when communications cross institu- tional or national boundaries. Authentication of platform-hosted content may require cooperation from platform operators, access to internal records, or compliance with foreign legal processes. By contrast, the cryptographic proof embedded within a provenance chain travels with the communica- tion itself. The content, signatures, and verification data remain available regardless of where the communication originated or which jurisdictions later become involved in the dispute. 7. Synthesis Existing evidentiary doctrine is well equipped to accommodate provenance chains. Rule 902(14)âs self-authentication framework for digitally identified data provides a natural mechanism for authenticating cryptographically signed digital speech acts, while traditional evidentiary tools remain available to establish the connection between public keys and particular persons. Challenges persistâincluding key attribution, claims of key compromise, timestamp accuracy, and the need to explain cryptographic verification to factfindersâbut these are familiar evidentiary questions rather than novel barriers to admissibility. The significance of provenance chains lies not merely in their technical reliability but in their evidentiary structure. Unlike platform-hosted content, whose authentication often depends on insti- tutional records and third-party cooperation, digital speech acts carry their own proof of integrity and provenance. Authentication therefore rests on cryptographic verification rather than the accuracy of a platformâs record-keeping systems. This evidentiary architecture reinforces the broader accountability framework discussed in this Part. Provenance chains provide a verifiable record of authorship and dissemination, making it pos- sible to identify who created content, who forwarded it, and in what sequence. Existing evidentiary rules can accommodate that record, enabling provenance information to be presented in court when disputes arise over ownership, attribution, or legal responsibility. E. International and Jurisdictional Complexity The grassroots architectureâs elimination of platform intermediaries does not eliminate jurisdic- tional questions; it redistributes them. Digital speech acts may originate in one country, be forwarded through participants in another, and cause alleged harm in a third. The resulting provenance chain creates a transparent record of authorship and dissemination, but it also raises difficult questions of personal jurisdiction, choice of law, and cross-border enforcement. This Section examines those questions and the legal uncertainties they create. 1. The Limits of a U.S.-Centric Analysis The analysis throughout this Part is primarily grounded in United States law. Discussions of Section 230, contributory copyright infringement, evidentiary authentication, personal jurisdiction, and defamation doctrine necessarily rely on U.S. statutes and precedent because those are the legal authorities governing the questions examined in this Article. The emphasis on U.S. law is also significant because many contemporary debates about digi- tal communication have been shaped by American doctrines governing intermediary liability and platform governance. Section 230 of the Communications Decency Act, for example, limits the circumstances under which online platforms and users may be treated as publishers of third-party content, while the Digital Millennium Copyright Act establishes safe harbors for qualifying service providers that facilitate online communications. 172 These frameworks have profoundly influenced the legal architecture of platform-mediated communication in the United States and provide an 172 47 U.S.C. § 230(c)(1); 17 U.S.C. § 512. 35 important baseline against which grassroots architectures can be evaluated. The underlying technological architecture is jurisdiction-neutral. A digital speech act carries the same cryptographic signature, provenance chain, and forwarding history whether created in New York, London, Singapore, or Berlin. What varies is the legal significance assigned to those facts. The distinction is important. Provenance chains do not determine liability; they determine attribution. They establish who created a digital speech act, who forwarded it, and when those actions occurred. Whether those facts give rise to defamation liability, copyright liability, regulatory exposure, or no liability at all depends on the substantive law applied by the relevant jurisdiction. For a United States audience, the most immediate questions concern how existing American doctrines would treat provenance-based communication systems. That inquiry provides a useful starting point because many of the legal questions discussed aboveâauthentication, authorship, intermediary liability, and personal jurisdictionâhave mature doctrinal frameworks in U.S. law. But the same provenance evidence could produce materially different outcomes under foreign legal systems that place different weight on speech rights, reputation interests, intermediary responsibility, or regulatory control. Accordingly, the analysis offered here should be understood as jurisdictionally bounded. The provenance mechanisms are global; the legal consequences are local. 2. Cross-Border Speech and Personal Jurisdiction When a person in Country A creates a digital speech act and someone in Country B forwards it to someone in Country C, which countryâs law applies? Which courts have jurisdiction over the creator and forwarders? Personal jurisdiction over digital speech. U.S. courts exercise personal jurisdiction over out- of-state defendants based on âminimum contactsâ with the forum such that the exercise of jurisdiction âdoes not offend traditional notions of fair play and substantial justice.â 173 For intentional torts like defamation, the âeffects testâ permits jurisdiction where a defendant expressly aimed conduct at the forum state and the brunt of the harm was suffered there. 174 In Calder, the Supreme Court upheld California jurisdiction over Florida defendants who wrote and edited a defamatory article about a California resident, because they knew the article would harm the plaintiff in California and intended to exploit the California market. 175 Applied to grassroots digital speech acts: If Alice in Country A creates defamatory content about Bob in Country B, Country B courts may have jurisdiction over Alice under Calder if she knew the content would harm Bob in Country B and expressly aimed the statement at Country B (e.g., by addressing it to Bobâs Country B community). But the Supreme Court has limited Calder: the defendant must engage in wrongful conduct âexpressly aimedâ at the forum, not merely cause foreseeable effects there. 176 This creates uncertainty for grassroots forwarders. If Carol in Country C forwards Aliceâs defamatory content to David in Country B, does Country B have jurisdiction over Carol? Under Walden, probably not, unless Carol expressly aimed the forwarding at Country B (knew David was in Country B, intended to harm Bob there, etc.). But if Carol forwards to a community spanning multiple countries, she may be subject to jurisdiction in any country where the content causes harmâa potentially global exposure. The amplification problem. Viral spread through grassroots forwarding creates cascading jurisdictional exposure. If Aliceâs content is forwarded through a chain spanning dozens of countries, each forwarder might face jurisdiction in the country where the original plaintiff resides, plus any countries where they forwarded the content. Unlike platformsâwhich face one global enforcement actionâgrassroots forwarders face potential jurisdiction wherever they directed content and wherever the plaintiff has connections. 173 International Shoe Co. v. Washington, 326 U.S. 310, 316 (1945). 174 Calder v. Jones, 465 U.S. 783, 789â90 (1984). 175 Id. 176 Walden v. Fiore, 571 U.S. 277, 287â90 (2014) (rejecting personal jurisdiction based solely on plaintiffâs forum contacts). 36 No simplified jurisdictional framework. The elimination of platform intermediaries makes personal jurisdiction more complex, not less. Each creator and forwarder must analyze whether their speech creates minimum contacts with foreign jurisdictions, whether the effects test applies, and whether their conduct was expressly aimed at the forum. This analysis must occur for every forwarding decision, not once by a central platform. 3. Choice of Law in Cross-Border Speech Even when a court has jurisdiction, which countryâs law applies to the speech? Defamation law varies dramatically by jurisdiction. U.S. defamation law provides robust protections for defendants: public figures must prove actual malice, truth is an absolute defense, and the First Amendment constrains liability. 177 English defamation law historically imposed stricter liability, placing the burden on defendants to prove truth and permitting larger damages. 178 Many countries criminalize defamation, creating exposure unavailable in U.S. civil law. When Alice in the U.S. creates content about Bob in the U.K., which law applies? If Carol in Germany forwards it to David in France, which law governs Carolâs liability? Traditional choice-of- law analysis examines: (1) where the plaintiffâs reputation was harmed (place of injury), (2) where the defendant acted (place of conduct), (3) which jurisdiction has the most significant relationship to the dispute, and (4) which jurisdictionâs law the parties expected to apply. 179 For digital speech forwarded globally, these factors point in multiple directions. The plaintiffâs reputation may be harmed in every country where the content appears. The defendant acted wherever they were physically located when forwarding. Multiple jurisdictions have relationships to the dispute. Parties rarely have clear expectations about which law applies to peer-to-peer forwarding. Forum shopping and the race to the bottom. Plaintiffs can sue in jurisdictions with the most favorable defamation laws. A public figure unable to prevail under U.S. actual malice standards might sue in a jurisdiction with stricter liability. The grassroots architecture provides no mechanism to prevent this forum shoppingâunlike platforms, which can sometimes invoke choice-of-law clauses in Terms of Service or argue for application of a single law globally. 4. Copyright and the Berne Convention Copyright presents somewhat more certainty due to international harmonization through the Berne Convention. 180 The Convention establishes minimum standards for copyright protection in over 180 member countries, including automatic protection upon fixation and national treatment (foreign works receive the same protection as domestic works). 181 Digital speech acts created by persons in Berne Convention countries receive copyright protection in all member countries. If Alice in France creates a digital speech act, it is automatically protected in the U.S., Japan, Australia, and all other Berne countries. This provides a degree of international copyright harmonization absent in defamation law. However, enforcement remains national and territorial. Copyright is a bundle of territorial rights, enforceable in each country separately. 182 If Bob in the U.S. forwards Aliceâs copyrighted digital speech act, Alice must sue Bob under U.S. copyright law in U.S. courts. If Carol in Germany also forwards it, Alice must sue Carol under German copyright law in German courts. There is no single global copyright enforcement actionâeach territorial infringement requires separate litigation. Practical enforcement challenges. A creator whose digital speech act is forwarded virally 177 See New York Times Co. v. Sullivan, 376 U.S. 254 (1964); Philadelphia Newspapers, Inc. v. Hepps, 475 U.S. 767 (1986). 178 But see Defamation Act 2013, c. 26 (U.K.) (reforming English defamation law to require claimants to prove serious harm). 179 See Restatement (Second) of Conflict of Laws § 145, 150 (1971). 180 Berne Convention for the Protection of Literary and Artistic Works, Sept. 9, 1886, as revised at Paris on July 24, 1971, and amended in 1979, S. Treaty Doc. No. 99-27 (1986). 181 Id. arts. 5(1)-(2). 182 See Subafilms, Ltd. v. MGM-Pathe Commcâns Co., 24 F.3d 1088, 1097 (9th Cir. 1994) (en banc) (copyright infringement is territorial). 37 across dozens of countries faces impossible enforcement burdens. They must: identify infringing forwarders in each country, retain counsel in each jurisdiction, navigate each countryâs copyright law and procedural rules, and obtain separate judgments in each forum. The grassroots architecture provides provenance chains identifying forwarders, but converting that technical information into legal enforcement across borders remains prohibitively expensive for most individual creators. 5. Regulation Without Platform Intermediaries A defining characteristic of contemporary internet regulation is its reliance on intermediaries. Legislatures frequently regulate online communications indirectly by imposing obligations on enti- ties that host, distribute, recommend, monetize, or otherwise facilitate user-generated content. This approach reflects a practical reality: platform operators aggregate communications that would other- wise be dispersed among millions of individual speakers and therefore function as regulatory leverage points. Rather than regulating each speaker individually, lawmakers often regulate the intermediary through which those speakers communicate. Many content regulations reflect this intermediary-focused model. The European Unionâs Digital Services Act requires platforms to implement content moderation systems, respond to takedown no- tices, assess systemic risks, and provide transparency reports. 183 Germanyâs Network Enforcement Act (NetzDG) similarly imposes obligations on social-media platforms to remove unlawful content within specified timeframes or face substantial fines. 184 Comparable approaches appear through- out modern platform governance, where compliance obligations are directed toward centralized intermediaries rather than toward individual speakers. Grassroots architectures remove that regulatory leverage point. Because content is stored on participant devices and disseminated through peer-to-peer forwarding, there is no platform operator capable of implementing moderation systems, processing takedown notices, producing transparency reports, or removing content globally. Existing intermediary-focused regulatory frameworks there- fore do not readily apply. If harmful content spreads through a grassroots network, there is no entity for regulators to hold accountable, no platform to fine for non-compliance, and no centralized moderation system to audit. Regulators would instead need to pursue individual speakers and forwardersâa practically impossible enforcement task when content spreads virally across thousands of participants. This creates a structural mismatch between intermediary-focused regulation and peer-to-peer communication. Content that would be swiftly removed from platforms (via regulatory pressure, takedown notices, or platform policies) may persist in grassroots networks because no single entity can remove it. Each forwarder must voluntarily stop forwarding, or be individually sued to compel removalâa process far slower and more resource-intensive than platform takedowns. This does not mean grassroots architectures exist outside legal regulation. Rather, the locus of regulation changes. Instead of imposing obligations on a platform intermediary, regulators must rely on existing legal doctrines directed at individual actors: civil liability, criminal sanctions, injunctions, and other remedies aimed at creators and forwarders themselves. The provenance chain may make such enforcement easier by identifying who created and disseminated particular content, but it does not solve the practical difficulties of pursuing large numbers of geographically dispersed participants. Grassroots architectures therefore replace centralized intermediary enforcement with decentralized actor-based enforcement. 6. Synthesis: Jurisdictional Complexity, Not Simplicity The grassroots architecture does not simplify international jurisdictional issuesâit redistributes them. Platform architectures centralize enforcement at a single chokepoint; grassroots architectures distribute exposure across all participants. This distribution has benefits (no single point of censor- ship, no platform gatekeeper) and costs (complex personal jurisdiction analysis for every speaker, forum shopping by plaintiffs, practically impossible enforcement for viral content). 183 Digital Services Act, arts. 14â16, 20, 23, 24, 34â35. 184 Gesetz zur Verbesserung der Rechtsdurchsetzung in sozialen Netzwerken [NetzDG] [Network Enforcement Act], Sept. 1, 2017, BGBl. I at 3352 (Ger.). 38 The choice between platform centralization and grassroots distribution is a choice between different jurisdictional paradigms: Platform model: Single entity subject to global enforcement, capable of one-action removal, vulnerable to regulatory pressure and forum shopping against the platform, provides predictable jurisdictional framework. Grassroots model: Distributed individuals with complex jurisdictional exposure, no single removal mechanism, resistant to centralized censorship, creates enforcement challenges for both plaintiffs and regulators. Which paradigm better serves free expression, accountability, and democratic values depends on contested normative judgments about the relative importance of enforcement capability versus resistance to centralized control. IX. ADDRESSING LEGAL CHALLENGES Part V argued that digital speech acts qualify as copyrightable works under Burrow-Giles v. Sarony, Feist v. Rural, and the Copyright Actâs fixation requirement. This Part addresses the strongest legal objections to that claim, engaging the doctrinal foundations and case law that challenge our thesis most directly. The objections fall into several categories: algorithmic generation challenges (whether machine- executed processes can produce copyrightable works), idea/expression distinctions (whether digital speech acts are unprotectable ideas or functional methods), merger doctrine (whether input-dictation prevents protection), mechanical generation precedents (cases denying copyright to outputs lacking human creativity), statutory limitations (contract preemption, first sale), and scholarly critiques (arguments that extending copyright to digital speech acts serves neither incentive nor democratic values). We address each in turn. A. Human Authorship and Algorithmic Execution The strongest contemporary challenge to copyrightability arises from the requirement of human authorship. Recent cases addressing artificial intelligence and non-human creators reaffirm that copyright protects only works attributable to human creative activity. In Thaler v. Perlmutter, the D.C. Circuit held that an artwork generated autonomously by an artificial intelligence system could not receive copyright protection because the work lacked a human author. 185 The court emphasized that âUnited States copyright law protects only works of human creation.â 186 Likewise, in Naruto v. Slater, the Ninth Circuit concluded that photographs taken by a macaque monkey could not be copyrighted because copyright law presupposes human authorship. 187 The U.S. Copyright Office has adopted a similar position in its recent guidance on artificial intelligence, explaining that copyright protection depends upon the presence of human creative control over the expressive elements of a work. 188 At first glance, digital speech acts appear to raise a similar concern. A person initiates a signing process, a cryptographic algorithm performs a series of computations, and a signature is generated automatically. If the resulting signature is produced by algorithmic operations rather than direct human manipulation, one might argue that the resulting work lacks the human authorship required by copyright law. The analogy is ultimately misplaced because it focuses on the mechanism that executes expression rather than the source of the expressive choices embodied in the work. The critical question under copyright law is not whether a machine participates in producing a work, but whether the expressive 185 Thaler v. Perlmutter, 687 F. Supp. 3d 140 (D.D.C. 2023), affâd, 130 F.4th 1039 (D.C. Cir. 2025). 186 Id. at 146. 187 Naruto v. Slater, 888 F.3d 418 (9th Cir. 2018). 188 U.S. Copyright Office, Copyright and Artificial Intelligence: Part 2 â Copyrightability (Jan. 2025), https://w.copyright.gov/ai/ai_part_2.pdf. 39 elements of that work originate in human decision-making. This was the central insight of Burrow- Giles. Faced with a photographic process that mechanically and chemically captured images, the Supreme Court did not ask whether the camera created the photograph. Instead, it examined whether the expressive features of the photograph originated with the photographer. Because Sarony selected the subject, pose, costume, lighting, and composition, the Court concluded that the resulting image embodied human authorship notwithstanding the cameraâs mechanical operation. 189 Digital speech acts present the same analytical structure. The person creating the digital speech act determines the expressive content of the work: what content to sign, when to sign it, what communicative force the act is intended to convey, and which identity to employ. These choices establish the meaning and significance of the expression. The cryptographic algorithm does not decide what is communicated, when it is communicated, or by whom it is communicated. It merely performs the mathematical operations necessary to bind those human choices to a verifiable signature. This distinction separates digital speech acts from the works at issue in Thaler and Naruto. In Thaler, the artificial intelligence system determined the expressive features of the output, while the human merely initiated the process. 190 In Naruto, the monkey determined the photographâs framing, angle, and moment of capture. 191 In both cases, the expressive choices that defined the resulting work originated with a non-human actor. Digital speech acts reverse that relationship. The human participant determines all expressive elements of the work, while the cryptographic system functions only as an execution mechanism. The analysis becomes clearer when the relevant work is properly identified. As argued in Part V, the copyrightable work is not the bare signature string viewed in isolation. It is the digital speech act as a whole: the expressive content, the cryptographic signature that authenticates it, and the metadata that situates it in time and identifies its speaker. The content is chosen by the person. The timing and identity reflected in the metadata are chosen or controlled by the person. The cryptographic signature serves to bind those elements together through mathematical authentication. The resulting work therefore embodies human expressive choices throughout, even though computational processes assist in executing and preserving them. This conclusion is consistent with the Copyright Officeâs distinction between AI-generated works and AI-assisted works. Copyright protection is unavailable where computational systems determine the expressive content of the work, but remains available where human beings exercise creative control and computational processes merely facilitate execution or implementation. 192 Digital speech acts fall squarely within the latter category. The algorithm authenticates expression; it does not create it. B. Idea, Expression, and Method Under Baker v. Selden Section 102(b) of the Copyright Act codifies a foundational principle of copyright law: copyright protects expression, not âany idea, procedure, process, system, method of operation, concept, prin- ciple, or discovery,â regardless of the form in which such matters are embodied. 193 The Supreme Court articulated this distinction in Baker v. Selden, holding that copyright in a book describing a bookkeeping system did not extend to the accounting method itself or to the forms necessary for its operation. 194 Seldenâs book explained a novel accounting method and included forms illustrating how the method was to be used. When Baker copied the forms, Seldenâs estate argued that copyright in the book extended to the accounting system embodied within it. The Supreme Court rejected that argument. Copyright protected Seldenâs explanation of the method, but not the method itself. As the Court explained, â[t]he description of the art in a book, though entitled to the benefit of copyright, lays no foundation for an exclusive claim to the art itself. The object of the one is explanation; the 189 Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53, 60 (1884). 190 Thaler, 687 F. Supp. 3d at 145â46. 191 Naruto, 888 F.3d at 420â21. 192 Copyright Office AI Guidance, supra note 188, at 9â13. 193 17 U.S.C. § 102(b). 194 Baker v. Selden, 101 U.S. 99 (1879). 40 object of the other is use.â 195 The significance of Baker is not merely that methods are excluded from copyright protection. Rather, the case establishes the analytical framework for distinguishing protected expression from the systems, procedures, and functional mechanisms through which expression may be created or communicated. Applied to digital speech acts, the relevant question is therefore not whether cryptographic signature systems are copyrightable. They are not. The question is whether a digital speech act constitutes protected expression created through the application of such a system, or whether it is merely an instance of the systemâs operation. The distinction becomes clearer when the work is properly characterized. Copyright does not protect the cryptographic signature algorithm itself. The algorithm is a mathematical procedure for authenticating content and therefore falls comfortably within §102(b)âs exclusion of procedures, processes, and methods of operation. 196 Nor does copyright protect the general concept of binding content to identity through cryptographic means. These functional and conceptual elements remain free for all to use. The digital speech act, however, is not the method. It is the expressive result produced through application of the method to particular human-authored content. In this respect, the relationship between a digital speech act and a cryptographic signature algorithm resembles the relationship between a photograph and a camera. Copyright does not protect the camera as a technology. It protects the photograph produced through the photographerâs creative choices. Likewise, copyright does not protect the signature algorithm. It protects the resulting expressive work when a person applies that method to content they have chosen, at a time they have chosen, and under an identity they have chosen to employ. This distinction also clarifies why digital speech acts differ from the forms at issue in Baker. The bookkeeping forms were unprotectable because they embodied the accounting method itself and contained no independent expressive content beyond the systemâs structure. 197 A blank ledger implemented the bookkeeping method but communicated nothing beyond that implementation. Protecting the forms would therefore have granted exclusive control over the method itself. Digital speech acts are fundamentally different. They contain expressive content beyond the authentication procedure. The text, image, audio, or video being signed is independently expressive and, where otherwise eligible, independently copyrightable. The signature does not replace that expression; it binds the expression to a particular person and moment in time. The resulting work therefore combines expressive content with cryptographic authentication in a manner that exceeds the mere operation of the underlying method. Modern software cases reinforce this distinction. In Lotus v. Borland, the First Circuit held that a menu command hierarchy constituted an uncopyrightable âmethod of operationâ because protecting it would effectively grant exclusive control over the way users operated the software. 198 Similarly, in Computer Associates v. Altai, the Second Circuit developed the abstractionâfiltrationâcomparison framework to distinguish protectable expression from unprotectable ideas, processes, methods, and functional requirements embedded within software. 199 Applying the logic of Altai to digital speech acts produces a straightforward result. At a high level of abstraction, a digital speech act implements the idea of authenticating content through a cryptographic signature. That idea is not copyrightable. The filtration step then removes the signature algorithm itself as an unprotectable mathematical method, the concept of identity binding as an unprotectable idea, and the authentication function as an unprotectable method of operation. What remains is the personâs expressive content, together with the deliberate choice to claim and authenticate that content under a particular identity at a particular moment. Those elements reflect human expressive choices rather than functional necessity. The Altai court emphasized that elements dictated by external functional requirements must be 195 Id. at 105. 196 See Lotus Dev. Corp. v. Borland Intâl, Inc., 49 F.3d 807, 815 (1st Cir. 1995) (holding menu command hierarchy was uncopyrightable method of operation), aff'd by an equally divided Court, 516 U.S. 233 (1996). 197 Baker, 101 U.S. at 104â07. 198 Lotus, 49 F.3d at 815. 199 Computer Assocs. Intâl, Inc. v. Altai, Inc., 982 F.2d 693, 706â11 (2d Cir. 1992). 41 filtered from copyright analysis. 200 Digital speech acts satisfy that requirement. The authentication function may dictate that a signature exists, but it does not dictate what content will be authenticated, when the authentication will occur, or which identity will be associated with it. Those choices originate with the person creating the digital speech act. The same analysis explains why the merger concerns underlying Baker are absent here. The bookkeeping forms in Baker merged with the accounting method because they were necessary to its use. 201 By contrast, copyright in a particular digital speech act does not prevent others from au- thenticating content, claiming authorship, or employing cryptographic signatures. Multiple persons remain free to create their own digital speech acts using the same methods and technologies. Pro- tecting one personâs signed expression therefore does not confer exclusive rights over the underlying ideas or methods. Properly understood, Baker v. Selden supports rather than undermines copyrightability. The case directs courts to distinguish expressive works from the systems through which they are produced. The cryptographic signature method remains outside copyrightâs domain. The digital speech act created through application of that method remains within it. C. Merger Doctrine and Input-Dictated Expression The merger doctrine prevents copyright protection when idea and expression become inseparable. Where there are so few ways to express an idea that protecting a particular expression would effectively grant exclusive rights over the idea itself, copyright protection must yield to §102(b)âs exclusion of ideas from the copyright monopoly. 202 The merger question arises in a distinctive form for digital speech acts. Unlike many expressive works, a cryptographic signature is computationally determined by its inputs. Given particular content, a particular private key, a particular cryptographic algorithm, and a particular moment of signing, only one valid signature can result. This raises the concern that the signature may be dictated by functional constraints in much the same way that certain software declarations or technical specifications are dictated by system requirements. 203 Properly framed, however, the merger inquiry does not ask whether a signature is mathematically determined once particular inputs are selected. The relevant question is whether protecting the resulting work would effectively monopolize the underlying idea. The doctrine is concerned with preserving public access to ideas, methods, and functional systems, not with denying protection whenever a technological process produces a determinate result. This distinction becomes clearer once the work is properly characterized. As discussed above, the copyrightable work is not the bare signature string viewed in isolation. It is the digital speech act as a whole: expressive content bound to identity and time through cryptographic authentication. The personâs choices regarding what to communicate, when to communicate it, and under which identity to claim it constitute the expressive elements of the work. The signature serves to authenticate those choices rather than replace them. The software cases illustrate the point. In Oracle, the method declarations at issue were dictated by the requirements of the Java programming language itself. 204 Programmers seeking to invoke par- ticular Java functions had no practical alternative but to employ the specified declarations. Protecting those declarations therefore threatened to restrict access to the underlying functionality. Digital speech acts present a fundamentally different situation. The expressive elements are not dictated by the authentication system. A person remains free to choose different content, different wording, different timing, and different identities. The fact that the resulting signature follows 200 Id. at 709â10. 201 Id. at 103. 202 See Morrissey v. Procter & Gamble Co., 379 F.2d 675, 678â79 (1st Cir. 1967) (â[W]hen there is essentially only one way to express an idea, the idea and its expression are inseparable and copyright is no bar to copying that expression.â). 203 See Oracle Am., Inc. v. Google Inc., 750 F.3d 1339, 1362â63 (Fed. Cir. 2014) (method declarations dictated by Java language requirements merged with idea), rev'd on other grounds, 141 S. Ct. 1183 (2021). 204 Oracle, 750 F.3d at 1362â63. 42 deterministically from those choices does not transform the choices themselves into functional requirements. The cryptographic process constrains only the mathematical implementation of the expression, not the expression itself. The distinction parallels the reasoning underlying Burrow-Giles. Once a photographer chooses a subject, pose, composition, lighting arrangement, and camera settings, the resulting image is produced through deterministic physical processes. 205 Yet copyright law has never regarded that determinism as a merger problem because the relevant expressive decisions originate with the author rather than with external functional constraints. The same principle applies here. The signature algorithm executes the authorâs choices through mathematical operations, but it does not dictate those choices. Nor does protection of a digital speech act threaten the sort of idea monopoly that merger doctrine is designed to prevent. Copyright in a particular signed expression does not prevent others from communicating the same underlying idea. It does not prevent others from authenticating content through cryptographic signatures. It does not prevent others from associating expression with identity or time. It prevents only the reproduction of a particular authenticated expression created by a particular person under particular circumstances. This point also distinguishes digital speech acts from the forms at issue in Baker. The bookkeeping forms merged with the accounting system because they were necessary to the operation of the system itself. 206 Protecting the forms would have effectively granted exclusive rights over the method. Digital speech acts do not occupy a comparable position. Protection of one personâs authenticated expression leaves the underlying authentication method entirely available for use by everyone else. Related concerns sometimes arise under the doctrine of scĂšnes Ă faire, which excludes elements that necessarily follow from a genre, setting, or subject matter. 207 Authentication systems do not dictate particular expressive content, particular moments of communication, or particular identities. Those elements remain matters of individual choice. The fact that a cryptographic signature must mathematically correspond to those choices does not render the choices themselves inevitable. The temporal and cryptographic uniqueness of digital speech acts further underscores the absence of merger. Even where identical content is signed repeatedly, each act of authentication occurs at a distinct moment and generates a distinct authenticated expression. Likewise, different persons authenticating identical content generate distinct expressions associated with distinct identities. Protection of one such expression therefore leaves the underlying ideas fully available to others through their own acts of communication and authentication. Merger doctrine therefore does not bar copyright in digital speech acts. The doctrine prohibits protection only when expression is so constrained that copyright would effectively confer ownership of an idea, method, or system. Digital speech acts do not create such a monopoly. The expressive choices embodied in the work remain distinct from the underlying concepts of authentication, attribution, and identity, and protection of those choices leaves the underlying ideas free for all to employ. D. Mechanical Reproduction and the Requirement of Originality A separate line of authority denies copyright protection to works produced through mechanical processes when those processes merely reproduce preexisting expression without the addition of human creative judgment. These cases are not principally concerned with technology. Rather, they address a more fundamental question: whether the purported author contributed any original expression beyond faithful reproduction of an existing work. The leading example is Bridgeman Art Library v. Corel Corp., which held that exact photographic reproductions of public-domain paintings lacked the originality necessary for copyright protection. 208 Although the photographs required technical skill, the objective was perfect fidelity to the underlying 205 Burrow-Giles, 111 U.S. at 60. 206 Baker, 101 U.S. at 103. 207 See Atari, Inc. v. North Am. Philips Consumer Elecs. Corp., 672 F.2d 607, 616 (7th Cir. 1982). 208 Bridgeman Art Library, Ltd. v. Corel Corp., 36 F. Supp. 2d 191 (S.D.N.Y. 1999). 43 paintings rather than the introduction of new expressive elements. As the court explained, where a photograph reproduces a public-domain work with complete accuracy, no new authorship is added. 209 The same principle appears in Meshwerks, Inc. v. Toyota Motor Sales U.S.A., Inc., where the Tenth Circuit denied copyright protection to highly detailed digital wire-frame models of Toyota vehicles. 210 The models required substantial labor and technical sophistication, yet they remained uncopyrightable because their purpose was exact replication rather than creative interpretation. The creators âdid not add any artistic interpretation or expression beyond what was evident in the original vehicles.â 211 Likewise, in ATC Distribution Group v. Whatever It Takes Transmissions & Parts, Inc., the Sixth Circuit treated product photographs designed to depict parts accurately for catalog purposes as lacking the originality required for copyright protection. 212 The court emphasized that photographs whose purpose is mechanical accuracy rather than expressive authorship do not satisfy copyrightâs originality requirement. 213 Taken together, these decisions establish a narrow but important proposition: copyright does not attach merely because a technologically sophisticated process produces an output. When the process serves only to reproduce an existing work with fidelity and leaves no room for meaningful creative judgment, the resulting output lacks the originality required by copyright law. Digital speech acts do not fall within this category. Unlike the works at issue in Bridgeman, Meshwerks, and ATC, digital speech acts are not reproductions of preexisting expression. The person creating the digital speech act determines the expressive content, the moment of publication, the identity under which the expression is asserted, and the communicative significance of the act itself. Those choices precede the operation of the cryptographic system and supply the originality embodied in the work. The cryptographic signature therefore performs a fundamentally different role from the technolo- gies considered in the mechanical-reproduction cases. In Bridgeman and Meshwerks, technology was employed to reproduce existing expression as accurately as possible. In digital speech acts, cryp- tographic computation does not reproduce expression; it authenticates expression already chosen by the human participant. The algorithm functions as a mechanism of attribution and verification rather than as a substitute for creative judgment. The distinction becomes clearer when contrasted with cases recognizing originality in technolog- ically mediated works. In Mannion v. Coors Brewing Co., the court emphasized that copyrightability depended upon the photographerâs creative decisions regarding lighting, composition, pose, and pre- sentation. 214 Distinguishing Bridgeman, the court observed that the photograph contained numerous original elements because it reflected the photographerâs expressive choices rather than a mere effort at faithful replication. 215 Digital speech acts resemble Mannion rather than Bridgeman. Their significance derives from human choices concerning content, timing, identity, and communicative purpose. The cryptographic process records and authenticates those choices, but it does not generate them. Moreover, it is not necessary to argue that the signature string, considered in isolation, possesses independent originality. As discussed above, the relevant work is the digital speech act as a whole: expressive content bound to a cryptographic signature and associated metadata. Originality resides in human-authored expression and human decisions embodied in the act. The signature contributes authentication and attribution, not independent creative content. Accordingly, the mechanical-reproduction cases do not undermine copyrightability. They deny protection where technology merely replicates preexisting expression without creative intervention. Digital speech acts involve the opposite relationship. Human authors make the expressive choices, 209 Id. at 197. 210 Meshwerks, Inc. v. Toyota Motor Sales U.S.A., Inc., 528 F.3d 1258 (10th Cir. 2008). 211 Id. at 1265â66. 212 ATC Distrib. Grp., Inc. v. Whatever It Takes Transmissions & Parts, Inc., 402 F.3d 700 (6th Cir. 2005). 213 Id. at 707. 214 Mannion v. Coors Brewing Co., 377 F. Supp. 2d 444 (S.D.N.Y. 2005). 215 Id. at 449. 44 while cryptographic systems serve only to preserve and verify those choices. The originality required by copyright law therefore remains attributable to the person rather than to the technology. E. Thin Copyright, Overclaiming, and the Limits of Protection The conclusion that digital speech acts satisfy copyrightâs requirements of authorship, originality, and fixation does not end the inquiry. A separate concern arises from the literature on âthin copyrightââthe proposition that works resting on only minimal originality may technically qualify for protection while nevertheless creating risks of overclaiming, anticommons effects, and restrictions on follow-on creation. Mark Lemley has argued that copyright protection for minimally creative works can produce forms of âthin copyrightâ that enable strategic overclaiming. 216 Copyright holders may attempt to leverage protection in genuinely creative elements to claim effective control over functional, factual, or otherwise unprotectable material, thereby generating anticommons problems that impede later creators. 217 Pamela Samuelson has similarly cautioned against applying copyrightâs minimal originality threshold so expansively that formulaic or functionally determined expressions receive protection beyond copyrightâs proper domain. 218 These concerns merit careful consideration because every digital speech act combines expression with authentication infrastructure. If copyright protection were recognized in such works, could it become a vehicle for extending proprietary control over authentication systems, network functions, or other functional elements of digital communication? The answer is no, both for architectural reasons and for doctrinal reasons. The architectural answer follows from the person-specific nature of digital speech acts. Thin copyright becomes problematic when protection over one work effectively constrains others from creating similar works or participating in a broader market. 219 Digital speech acts do not operate in this manner. Copyright in one personâs authenticated expression does not prevent others from creating their own authenticated expressions, even where the underlying content is identical. Each cryptographic signature is associated with a distinct identity and a distinct act of authentication. Protection therefore remains inherently person-specific rather than market-wide. Nor does recognition of copyright in digital speech acts threaten the sort of bundling problem that concerns Lemley. The functional components of the system remain analytically distinct from the expressive components. The cryptographic algorithms, authentication procedures, verification processes, and key-management systems remain unprotectable methods and processes. Copyright attaches, if at all, only to the resulting expression created through their use. 220 Recognition of copyright in the work therefore does not convert the underlying authentication architecture into proprietary subject matter. The same distinction substantially limits any danger of restricting follow-on creation. 221 Recogni- tion of copyright in Aliceâs digital speech act does not prevent Bob from creating his own authenticated expression. It does not prevent third parties from developing software that creates, processes, dis- plays, indexes, transmits, or verifies digital speech acts. It does not prevent commentary, quotation, criticism, or other uses governed by traditional copyright limitations. Nor does it prevent analysis of the factual information contained within digital speech acts. The functional and informational dimensions of the system remain available to others even if particular expressions receive protection. Grassroots architectures further reduce the possibility of strategic overclaiming. Because digital speech acts remain under the control of the individuals who create them, no centralized repository accumulates large portfolios of minimally creative works capable of being deployed to control 216 Lemley, supra note 128, at 1059â65. 217 Id. 218 Samuelson, supra note 129. 219 See Feist Publâns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 349â50 (1991) (rejecting such claims for factual compilations). 220 Lemley, supra note 128, at 1062â65. 221 Id. at 1065â72. 45 markets or exclude competitors. Copyright ownership remains distributed in the same manner as the communications themselves. Even if these architectural features were disregarded, existing copyright doctrine independently constrains the scope of any protection that digital speech acts might receive. As Part IX.C demon- strated, merger doctrine denies protection where idea and expression become inseparable. 222 El- ements dictated by functional requirements or computational necessity remain outside copyrightâs reach regardless of how generously originality is construed. Any aspect of a digital speech act that is entirely determined by its inputs is therefore filtered from protection through ordinary merger analysis. Section 102(b) provides a second and independent limitation. The statute excludes from copy- right protection âany idea, procedure, process, system, method of operation, concept, principle, or discoveryâ regardless of the originality embodied in surrounding expression. 223 The authentication functions of cryptographic systems, the procedures through which signatures are generated and veri- fied, and the broader architecture of identity binding therefore remain categorically outside the scope of copyright. These exclusions operate regardless of whether a court regards the expressive aspects of digital speech acts as minimally or substantially creative. The significance of these doctrinal limitations is that they prevent thin copyright from expanding into the forms of strategic overclaiming identified by Lemley and Samuelson. Merger doctrine removes protection from input-dictated elements. Section 102(b) removes protection from methods, systems, and processes. What remains after those doctrines operate is the personâs particular expressive act of communication, attribution, and authentication. Accordingly, recognizing copyright in digital speech acts does not produce the anticommons concerns commonly associated with thin copyright. Both the architecture of grassroots systems and the internal structure of copyright doctrine confine protection to the expressive choices that justify copyright in the first place, while leaving the underlying methods of authentication, communication, and network participation available for all to use. F. Remaining Statutory and Policy Objections 1. Contract Preemption Under Section 301 Section 301 preempts state-law rights that are equivalent to copyrightâs exclusive rights. 224 Preemption applies when a claim concerns subject matter within copyright and seeks to enforce rights equivalent to those granted by §106. 225 The central inquiry is not whether a contractual obligation concerns copyrighted works, but whether the obligation merely duplicates copyright or instead rests on an independent legal basis. Courts generally resolve this question through the âextra elementâ test. A state-law claim survives preemption when it requires proof of an element beyond those necessary to establish copyright infringement. 226 Copyright infringement requires unauthorized exercise of one of the exclusive rights granted by §106. Contract liability, by contrast, ordinarily depends upon proof of agreement, breach, and resulting injury. Because contractual obligations arise from voluntary assent rather than from the mere act of copying or distributing a work, courts have consistently treated contractual claims as containing the requisite extra element. 227 Digital social contracts fall comfortably within this framework. The non-unbundling obligations described in Part I do not arise from copyright ownership itself. They arise from agreements among participants concerning the treatment of digital speech acts within a particular network. A participant who strips signatures, destroys provenance information, or falsifies attribution violates 222 See Morrissey v. Procter & Gamble Co., 379 F.2d 675, 678â79 (1st Cir. 1967). 223 17 U.S.C. § 102(b). 224 17 U.S.C. § 301(a). 225 Id. 226 See Computer Assocs. Intâl, Inc. v. Altai, Inc., 982 F.2d 693, 716 (2d Cir. 1992). 227 See ProCD, Inc. v. Zeidenberg, 86 F.3d 1447, 1454â55 (7th Cir. 1996) (contract claim not preempted because it requires proof of extra elementâagreementâbeyond copyright infringement). 46 those obligations because the participant agreed not to do so, not merely because the participant copied or distributed a copyrighted work. The requirement of assent therefore supplies the extra element that takes such claims outside §301 preemption. 228 The interests protected by digital social contracts also differ from the interests protected by copy- right. Copyright secures exclusive rights in reproduction, distribution, adaptation, performance, and display. Digital social contracts instead govern attribution, provenance preservation, and crypto- graphic integrity. A participant may comply fully with copyright law while nonetheless violating a contractual obligation not to remove signatures or disrupt provenance chains. Conversely, a par- ticipant might infringe copyright without breaching any contractual obligation. The two regimes therefore regulate different conduct and vindicate different interests. This distinction reflects a broader difference between copyright and contract. Copyright creates rights enforceable against the world. Contract creates obligations that bind only those who agree to them. As Judge Easterbrook observed in ProCD, copyright establishes rights against strangers, whereas contracts generally affect only their parties. 229 Digital social contracts operate in precisely this manner. They do not create exclusive rights in expression beyond those already recognized by copyright law; they establish network-specific obligations governing how participants will treat authenticated communications. Attribution obligations warrant brief separate consideration because copyright law itself recog- nizes limited attribution and integrity interests through the Visual Artists Rights Act. 230 VARA, however, applies only to a narrow category of works of visual art. 231 Digital speech acts generally fall outside that category. Contractual commitments to preserve signatures and provenance therefore do not duplicate the attribution rights Congress created through VARA. Moreover, attribution-related claims frequently involve additional elements such as deception, fraud, or false designation, which further distinguish them from ordinary copyright claims. 232 Digital social contracts and copyright thus perform complementary rather than equivalent func- tions. Copyright establishes ownership of digital speech acts and provides remedies against unau- thorized exploitation by third parties. Digital social contracts establish consensual rules governing attribution, provenance, and non-unbundling within particular communities. Because those obliga- tions depend upon agreement and protect interests distinct from the exclusive rights enumerated in §106, they are not preempted by §301. 233 2. First Sale Doctrine and the Limits of Section 109 Section 109(a) provides that âthe owner of a particular copy or phonorecord lawfully made under this title . . . is entitled, without the authority of the copyright owner, to sell or otherwise dispose of the possession of that copy or phonorecord.â 234 The first sale doctrine thus limits a copyright ownerâs ability to control downstream transfers of a lawfully acquired copy. Once ownership of a particular copy has passed, the copyright owner may not invoke the distribution right to prevent that copy from being resold, lent, or otherwise transferred. At first glance, the doctrine appears to create tension with the non-unbundling norms described in Part I. If a recipient lawfully acquires a copy of a digital speech act, one might argue that first sale permits the recipient to redistribute the work in whatever form they choose, including by removing signatures, severing provenance information, or otherwise altering the relationship between content and attribution. That conclusion does not follow from the structure of §109. First sale limits only certain copyright interests associated with a particular copy. It does not confer a general right to alter a copyrighted 228 See Computer Assocs., 982 F.2d at 716; ProCD, 86 F.3d at 1454â55. 229 ProCD, 86 F.3d at 1454. 230 17 U.S.C. § 106A. 231 17 U.S.C. § 101 (defining âwork of visual artâ). 232 See Dastar Corp. v. Twentieth Century Fox Film Corp., 539 U.S. 23 (2003) (Lanham Act false designation claim involves elements beyond copyright). 233 See Computer Assocs., 982 F.2d at 716; ProCD, 86 F.3d at 1454â55. 234 17 U.S.C. § 109(a). 47 work, create new versions of it, or disregard independent contractual obligations governing its use. Most fundamentally, §109 concerns the distribution of a lawfully acquired copy, not the creation of new copies or new works. The doctrine permits disposition of âthat copyâ which the recipient owns. 235 It does not authorize conduct that implicates rights beyond distribution. A recipient therefore may transfer the digital speech act as received, including its associated signatures and provenance information. What §109 does not obviously authorize is the creation and redistribution of a modified version from which those elements have been removed. This distinction is particularly important because the copyright claim advanced throughout this Article treats the digital speech act as the relevant work. The work consists not merely of semantic content in isolation, but of the content as authenticated, attributed, and temporally situated through the act of signing. Removing those elements does not simply transfer an existing copy; it creates a different object than the one originally received. To the extent such conduct constitutes reproduction or preparation of a modified work, it implicates rights that first sale does not extinguish. 236 The difficulty becomes even more pronounced in digital environments, where redistribution ordinarily occurs through copying rather than through physical transfer of an existing object. The traditional first-sale doctrine developed in a world of books, records, and paintings, where possession of a copy could pass from one owner to another without reproduction. Digital communication generally operates differently. The Second Circuitâs decision in Capitol Records, LLC v. ReDigi Inc. illustrates the point. 237 ReDigi attempted to create a secondary market for digital music files by transferring usersâ purchased songs to new buyers. The court held that the process fell outside first-sale protection because digital transfer necessarily involved reproduction rather than mere distribution of the original copy. 238 The doctrine protected disposition of a particular copy; it did not authorize the creation of a new one. Digital speech acts present a similar problem. Forwarding a signed expression ordinarily creates additional copies on additional devices rather than transferring possession of a single physical artifact. To the extent redistribution requires reproduction, the conduct falls outside the traditional scope of §109 even before questions of attribution or provenance arise. 239 Nor does first sale displace the contractual dimension of grassroots architectures. The doctrine limits copyright claims; it does not invalidate contractual obligations voluntarily assumed by partici- pants. This distinction has long been recognized in copyright jurisprudence. Although Bobbs-Merrill Co. v. Straus established the first-sale doctrine as a limit on copyright ownersâ ability to control downstream resale, it did not eliminate the possibility of contractual restrictions operating indepen- dently of copyright law. 240 Likewise, ProCD, Inc. v. Zeidenberg recognized that contractual use restrictions may coexist with copyright limitations because contract obligations arise from agreement rather than from exclusive rights enforceable against the world. 241 Digital social contracts operate in precisely this manner. Their purpose is not to expand copy- rightâs exclusive rights but to establish mutually accepted obligations concerning attribution, prove- nance preservation, and non-unbundling. A participant who agrees not to strip signatures may remain bound by that commitment even where copyright law alone would not prohibit a particular transfer. The source of the obligation is contractual and architectural rather than proprietary. Indeed, the non-unbundling norm is enforced primarily through technical design and social agreement rather than through copyright litigation. Grassroots applications can validate signatures, preserve provenance chains, reject altered records, and present authenticated content in ways that make attribution-preserving transmission the default mode of communication. Copyright functions largely as a legal backstop protecting the integrity of the underlying digital speech act; the day-to-day enforcement of attribution norms occurs through architecture and contract. Accordingly, first sale does not undermine the theory of copyright advanced here. Section 109 235 17 U.S.C. § 109(a) (emphasis added). 236 See 17 U.S.C. § 109(a) (preserving rights beyond the limited transfer privilege recognized by the doctrine). 237 Capitol Records, LLC v. ReDigi Inc., 910 F.3d 649, 657â60 (2d Cir. 2018). 238 Id. 239 Id. 240 Bobbs-Merrill Co. v. Straus, 210 U.S. 339, 350â51 (1908). 241 ProCD, Inc. v. Zeidenberg, 86 F.3d 1447, 1454 (7th Cir. 1996). 48 permits redistribution of lawfully acquired copies, but it does not create a right to sever authorship information, eliminate provenance, reproduce works during transmission, or disregard contractual commitments concerning attribution. The doctrine limits copyrightâs distribution right, yet leaves intact both the contractual obligations and the architectural mechanisms through which grassroots systems preserve the integrity of digital speech acts. 3. Scholarly Critiques and the Scope of Copyright Copyright scholars have long cautioned against extending copyright beyond its traditional domain. Jessica Litman, James Boyle, and others have argued that many modern expansions of copyright have produced additional private rights without corresponding public benefits, often serving incumbent interests at the expense of access, competition, and follow-on creation. 242 From this perspective, recognizing copyright in digital speech acts might appear to be another step in copyrightâs continual expansion into new technological environments. 243 That characterization, however, depends on treating digital speech acts as a new category of subject matter requiring doctrinal enlargement. The argument advanced here is narrower. Digital speech acts are not proposed as a new class of copyrightable works. Rather, they are expressions that already satisfy the statutory requirements of originality, authorship, and fixation discussed in Part V. Copyright protection therefore arises not from expanding copyrightâs boundaries but from applying existing doctrine to expressions created within a new technical architecture. 244 The incentive critique raises a related objection. Copyrightâs traditional justification is utilitarian: exclusive rights are granted to encourage creative production. 245 Yet individuals routinely create messages, photographs, comments, and social-media posts without regard to copyright incentives. Digital expression appears abundant even in the absence of exclusive rights. 246 The force of that critique depends upon viewing copyright exclusively as an incentive to create expressive content. Copyright, however, has long been understood to serve additional values, includ- ing attribution, integrity, control over commercial exploitation, and protection against unauthorized appropriation. 247 In the context of digital speech acts, the principal incentive is not necessarily the production of additional expression. Rather, copyright helps support participation in architectures where speakers retain ownership of their own expressions instead of transferring effective control to centralized intermediaries. The incentive at issue is therefore architectural as much as creative. A different scholarly concern focuses on formality. Historically, copyright law relied on registra- tion, notice, and related formalities that distinguished works whose authors sought legal protection from ordinary communication. 248 Automatic copyright protection risks granting legal rights to works whose creators neither need nor desire protection. 249 Digital speech acts partially answer this concern through architecture. Participation in a crypto- graphically authenticated system requires deliberate acts of identification and signing. The creation of a digital speech act is therefore not entirely analogous to casual speech or ephemeral communica- tion. Moreover, grassroots systems can incorporate licensing preferences, attribution requirements, and provenance rules directly into their operation, creating technical signals that function in ways similar to historical copyright formalities while remaining consistent with modern copyright law. Contemporary digital platforms already employ analogous mechanisms through account authentica- tion, copyright-management systems, notice-and-takedown procedures, content-identification tools, 242 Jessica Litman, Digital Copyright (2001); James Boyle, The Public Domain: Enclosing the Commons of the Mind (2008). 243 Boyle, supra, at 54â97. 244 17 U.S.C. § 102(a). 245 See U.S. Const. art. I, § 8, cl. 8 (granting Congress power â[t]o promote the Progress of Scienceâ through exclusive rights). 246 See Yochai Benkler, Sharing Nicely: On Shareable Goods and the Emergence of Sharing as a Modality of Economic Production, 114 Yale L.J. 273 (2004). 247 See Neil Weinstock Netanel, Copyrightâs Paradox 53â82 (2008) (copyright serves multiple democratic functions beyond incentive). 248 See Christopher Sprigman, Reform(aliz)ing Copyright, 57 Stan. L. Rev. 485 (2004). 249 Id. at 490â505. 49 and metadata-based attribution practices. Grassroots architectures do not introduce formalization into digital expression so much as relocate these functions from centralized intermediaries to the users who create and control the underlying works. Platform-power critiques likewise require careful attention. Copyright in user-generated content has often benefited intermediaries more than speakers because platforms obtain broad licenses or transfers through Terms of Service and then exploit those rights at scale. 250 If digital speech acts merely generated additional rights that could be appropriated by platforms, the critique would have considerable force. The architecture examined in this Article, however, reverses that dynamic. Copyright ownership remains attached to individual speakers, while possession of the relevant files remains distributed among participants rather than centralized platforms. Rather than facilitating appropriation, copy- right in digital speech acts strengthens the legal position of individual speakers against intermediaries. The concern that copyright expansion inevitably serves concentrated interests is therefore less per- suasive where ownership itself remains decentralized. 251 Lastly, scholars have expressed concern that copyright protection for political and social ex- pression may burden democratic discourse. 252 Copyrightâs application to slogans, protest speech, commentary, and movement communications can appear difficult to reconcile with robust democratic participation. Existing doctrine, however, already contains substantial safeguards. Fair use protects criticism, commentary, and transformative engagement with copyrighted expression. 253 The idea- expression distinction and fair-use doctrine together form part of the constitutional accommodation between copyright and free expression. 254 Within grassroots architectures, copyright protection primarily supports attribution and resistance to unauthorized appropriation while leaving substantial space for circulation, criticism, and democratic engagement. These scholarly critiques identify genuine concerns, but they do not establish that copyright is inappropriate for digital speech acts. Much of the skepticism surrounding copyright expansion assumes centralized ownership, intermediary control, and concentration of rights. Digital speech acts operate within a markedly different architecture in which ownership, possession, and authorship remain aligned with individual speakers. Whether that alignment ultimately produces better social outcomes remains an empirical question. The doctrinal question addressed in this Article is narrower: existing copyright principles are capable of recognizing digital speech acts without generating the forms of concentration, appropriation, and restriction that motivate many contemporary critiques of copyright expansion. G. From Copyrightability to Property and Self-Governance Part V established the affirmative case that digital speech acts satisfy copyrightâs traditional re- quirements of authorship, originality, and fixation. 255 This Part examined the principal doctrinal and policy objections to that conclusion, including algorithmic generation, idea-expression limitations, mechanical reproduction doctrine, merger, thin copyright concerns, contract preemption under §301, first sale under §109, and broader scholarly critiques of copyright expansion. Taken together, however, these objections do not displace the central conclusion of this Article. Digital speech acts remain expressions created through human volition, fixed in tangible media, and sufficiently original to satisfy copyrightâs low threshold for protection. The cryptographic mechanisms through which they are authenticated do not eliminate authorship, transform expression into process, or convert creative acts into purely mechanical reproductions. 256 Existing doctrines 250 See Tarleton Gillespie, Wired Shut: Copyright and the Shape of Digital Culture 137â69 (2007). 251 See Boyle, supra, at 54â77 (documenting copyright expansion benefiting media companies). 252 See Rebecca Tushnet, Copy This Essay: How Fair Use Doctrine Harms Free Speech and How Copying Serves It, 114 Yale L.J. 535 (2004). 253 See Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569 (1994). 254 See Eldred v. Ashcroft, 537 U.S. 186, 219â20 (2003) (copyrightâs built-in free speech protections through idea/expression and fair use). 255 See Part V. 256 See Parts IX.A,B, D. 50 such as merger and §102(b) continue to limit protection where functionality predominates 257 , while doctrines governing thin copyright, preemption, and first sale constrain the scope of any rights recognized. The result is not an expansion of copyright beyond its traditional boundaries, but an application of established principles to a new technological environment. This conclusion does not resolve every question. Courts may disagree about the significance of cryptographic authentication, the level of creativity required for particular digital speech acts, or the broader policy implications of recognizing copyright in decentralized communications. Reasonable disagreement remains possible. The relevant inquiry, however, is not whether uncertainty exists, but whether existing doctrine forecloses recognition of copyright in digital speech acts. It does not. Copyrightability alone, however, is not the Articleâs endpoint, but its legal foundation. The question addressed in this Part has been whether existing doctrine permits recognition of copyright in digital speech acts. The answer, as argued above, is yes. Yet the significance of that conclusion lies not merely in the recognition of rights, but in the legal relationships such recognition makes possible. Where ownership, authorship, attribution, and possession become capable of coalescing within the same expressive act, copyright ceases to function solely as a mechanism of exclusion and instead becomes relevant to broader questions concerning property, participation, and governance. The implications of that possibility extend beyond copyright doctrine itself and into the institutional arrangements through which digital communities organize, coordinate, and govern themselves. X. PROPERTY AND DEMOCRATIC SELF-GOVERNANCE Political theorists across traditions have long recognized that property is not merely an economic institution but a democratic one. The central concern is not wealth accumulation as such, but independence. Citizens who possess productive assets participate in social, economic, and political life on different terms than citizens whose participation depends upon the continuing permission of others. Ownership provides a sphere of autonomy within which individuals can act, create, associate, and speak without being wholly subject to external control. A. From Labor to Democratic Independence John Locke grounded property rights in self-ownership and labor. 258 One owns oneself and therefore acquires claims in the products of oneâs labor. 259 Although Lockean labor theory remains influential, its application to intellectual property has been extensively criticized. Scholars have questioned whether labor alone can justify exclusion rights in non-rivalrous goods, whether creators can claim ownership over expressions built from common cultural resources, and whether intellectual property creates artificial scarcity rather than protecting genuine interests. 260 Those critiques substantially weaken any attempt to justify digital speech act ownership solely through labor theory. The democratic case developed here therefore rests less on labor-desert than on a broader concern with independence, dispersed power, and resistance to domination. The question is not simply whether creators deserve ownership because they labored. The more fundamental question is whether democratic participation in digital society requires individuals to own the expressive assets through which they communicate, associate, and create value. This concern becomes particularly salient in digital environments. Contemporary social, polit- ical, and economic participation increasingly occurs through digital expression. Individuals build 257 Parts IX.C 258 Locke, supra note 24, § 25â51. 259 Id. § 27. 260 Edwin C. Hettinger, Justifying Intellectual Property, 18 Phil. & Pub. Aff. 31, 35â38 (1989); Wendy J. Gordon, A Property Right in Self-Expression: Equality and Individualism in the Natural Law of Intellectual Property, 102 Yale L.J. 1533, 1540â49 (1993); Justin Hughes, The Philosophy of Intellectual Property, 77 Geo. L.J. 287, 296â330 (1988); Seana Valentine Shiffrin, Lockean Arguments for Private Intellectual Property, in New Essays in the Legal and Political Theory of Property 138 (Stephen R. Munzer ed., 2001). 51 reputations, form associations, organize collective action, create economic value, and engage in public discourse through digital communications. If ownership of those expressions is concentrated in intermediaries, participation itself becomes dependent on intermediary permission. If ownership remains with the individuals who create the expressions, participation occurs on a more independent footing. The democratic significance of ownership therefore lies not primarily in rewarding labor, but in structuring relationships of power. Ownership determines who controls participation, who captures value, and who possesses the legal authority to decide how digital expressions may be used. It is this question of distributed versus concentrated control that links theories of property to theories of democratic self-government. Existing debates frequently ask how individuals can retain greater control over information collected by platforms. This Article asks a prior question: who owns the productive assets generated through digital participation in the first place? Privacy concerns arise after ownership has already been allocated. Democratic theory requires examination of the allocation itself. B. Digital Speech Acts as Productive Assets Much contemporary scholarship concerning digital ownership frames ownership as a mechanism for protecting privacy, preserving autonomy, or increasing user control over digital systems. These concerns are important, but they remain largely reactive. They seek to constrain the power of existing intermediaries without addressing the underlying distribution of productive assets that gives rise to intermediary power in the first place. The argument advanced here is different. Digital expressions are not merely objects of privacy or autonomy interests. They increasingly function as productive assets through which individuals generate economic value, social capital, political influence, and participation in public life. The democratic question is therefore not simply whether users retain control over their information, but whether ownership of these productive assets remains distributed among people or becomes concentrated within platform institutions. Among contemporary democratic theories of property, Rawlsâs account of property-owning democracy provides the strongest framework for understanding ownership of digital speech acts. 261 Rawls distinguished property-owning democracy from welfare-state capitalism on the basis of ownership. In welfare-state capitalism, productive assets remain concentrated in relatively few hands, while the majority participate primarily as employees or beneficiaries of redistribution. In a property- owning democracy, productive assets are widely dispersed throughout society, enabling citizens to participate as independent agents rather than dependents. 262 As Rawls explained, the objective is to prevent a small segment of society from controlling economic life and thereby exercising indirect control over political life as well. 263 The central insight is structural. Democratic equality requires more than formal rights. It requires sufficiently dispersed ownership that citizens can participate in social and economic life without being subject to the arbitrary power of concentrated private institutions. Although Rawls wrote primarily with land, capital, and productive enterprises in mind, the under- lying principle is not limited to physical assets. The relevant category is productive assetsâresources that enable meaningful participation in the economy and society. 264 In a digital society, expressive and informational assets increasingly function as productive assets. Digital expressions generate reputation, social capital, economic opportunity, political influence, and community participation. Ownership of those expressions therefore affects whether individuals participate in digital life as independent actors or merely as tenants within privately owned platforms. 261 Rawls, supra note 25, at 135â40. 262 Id. at 139. 263 Id. 264 See Martin OâNeill & Thad Williamson, Property-Owning Democracy: Rawls and Beyond (2012); Samuel Freeman, Capitalism in the Classical and High Liberal Traditions, 28 Soc. Phil. & Polây 19 (2011); Thad Williamson, Realizing Property-Owning Democracy: A 20-Year Strategy to Create an Egalitarian Distribution of Assets in the United States, in OâNeill & Williamson, supra, at 225. 52 The classification of digital expressions as productive assets is not theoretical speculation. Con- temporary platform firms derive extraordinary economic value from user-generated content and associated social interactions. As discussed in Part I, public filings from Meta and Alphabet reveal revenues measured in the hundreds of billions of dollars generated through systems built upon user participation, expression, and engagement. The economic significance of these activities demon- strates that digital expression has become an important productive resource within contemporary society. The question for democratic theory is therefore not whether such assets exist, but whether ownership of them remains broadly distributed among the individuals who create them or becomes concentrated within the institutions that aggregate and monetize them. Rawlsâs concern was never with any particular form of capital. His concern was with the democratic consequences of concentrated ownership. In agrarian societies the relevant assets were land; in industrial societies they were factories and machinery. In digital societies, expressive and informational assets increasingly perform the same productive function. If ownership of those assets becomes concentrated within a small number of platform intermediaries, the structural conditions Rawls identified reappear in digital form. This perspective reframes copyright ownership of digital speech acts. The question is not whether a single post, image, or message possesses substantial economic value. The question is whether the aggregate ownership of digital expression is distributed among citizens or concentrated within intermediary institutions. Platform architectures concentrate control over digital participation. Platforms own the infrastructure, control access to audiences, impose terms of service, and frequently acquire broad licensing rights in user-generated content. Users participate, but the productive assets underlying participation remain institutionally controlled. Grassroots architectures reverse that arrangement. Individuals possess the devices, control the keys, retain the copyrighted expression, and maintain possession of the data itself. Ownership of productive digital assets is dispersed rather than concentrated. Yet ownership alone does not fully capture the democratic significance of grassroots architectures. What distinguishes these systems is not merely the distribution of ownership, but the reunification of ownership and possession. C. Ownership, Possession, and Distributed Power The democratic significance of ownership is magnified in grassroots architectures because own- ership and possession coincide. Unlike platform environments, where users may create expressions while platforms retain possession of the underlying infrastructure and often acquire broad rights of use, grassroots systems place both possession and ownership in the hands of participants themselves. The combination of cryptographic possession and copyright ownership creates a form of digital independence that neither technical control nor legal rights could achieve alone. From a Rawlsian perspective, this dispersion is not merely economically desirable. It advances the democratic objective of preventing domination through concentrated ownership. The copyrightability of digital speech acts therefore functions as one mechanism through which ownership of productive assets can remain broadly distributed in an increasingly digital economy. The concentration of ownership matters because ownership structures are ultimately governance structures. As Joseph Singer argues, property is not merely a relationship between persons and things, but a structure of relationships among persons. 265 Ownership therefore determines who possesses power over whom. Similar observations appear in scholarship on digital governance. As Luca Belli and Jamila Venturini observe, one of the primary functions of contracts is to fashion power relationships. 266 Platform Terms of Service therefore function not merely as agreements but as instruments that allocate authority within digital environments. They determine who may speak, who may access audiences, who may monetize expression, and who may exclude others from participation. Property and contract thus operate together to structure participation in digital life. Platform ownership structures create relationships of dependence because platforms control both the productive assets of participation and the contractual conditions under which participation 265 Singer, supra note 26. 266 Luca Belli & Jamila Venturini, Private Ordering and the Rise of Terms of Service as Cyber-Regulation, 5 Internet Policy Rev. 1, 3 (2016). 53 occurs. Grassroots architectures redistribute these relationships by returning both possession and ownership to participants themselves, reducing dependence upon intermediary institutions and dis- persing authority throughout the network. The result is not the elimination of governance, but the decentralization of governance authority itself. D. Democratic Ownership and Anti-Enclosure Commons-oriented scholars may object that democratic participation is better served through shared informational commons than through proprietary rights. That concern is important, partic- ularly given the history of copyright expansion and enclosure of information resources. Yet the alternative examined in this Article is not a choice between commons and ownership in the abstract. It is a choice between ownership concentrated in platform intermediaries and ownership distributed among participants. Democratic alternativesâincluding commons-based governance, open licens- ing, cooperative production, and other forms of collective stewardshipâdepend upon participants retaining meaningful authority over the expressive assets they create. Grassroots architectures pre- serve those possibilities by ensuring that ownership remains with creators rather than intermediary institutions. Participants may choose to share, license, pool, or collectively govern their expressive assets, but those decisions remain theirs to make. The democratic concern is therefore not enclosure of the commons by individuals, but enclosure of the digital commons by platforms. In this respect, copyright serves an anti-enclosure function. The right is not invoked to cen- tralize control over expression, but to prevent the transfer of expressive ownership from creators to intermediary institutions. No single theory fully justifies ownership of digital speech acts. Lockean labor theory faces familiar objections. Rawls did not write about copyright. Singerâs relational theory imposes obli- gations as well as rights. Commons scholars correctly warn against enclosure. Yet these traditions converge on a common concern: preventing domination through concentrated control of the re- sources necessary for participation. In digital society, expressive assets increasingly constitute such resources. Copyright ownership of digital speech acts thus serves democratic values not because it maximizes exclusion, but because it prevents the consolidation of expressive ownership within intermediary institutions. It disperses control, preserves attribution, and provides a legal foundation for digital participation independent of platform permission. The remaining question is practical rather than theoretical: if ownership can be distributed in this manner, what consequences follow for governance, economic organization, and the operation of digital networks themselves? The next Part turns to those institutional implications. XI. ECONOMIC CONSIDERATIONS The democratic significance of ownership is matched by its economic significance. Ownership of digital speech acts does not merely allocate control over participation; it allocates the capacity to capture economic value generated through participation. Where platform architectures concentrate this capacity within intermediary firms, grassroots architectures return it to participants themselves. As discussed in Part I, contemporary platform firms derive extraordinary revenues from systems built upon user expression, interaction, and engagement. The economic question is therefore not merely how digital value is created, but how ownership of the assets generating that value is allocated. Under platform architectures, users create expressive content while platforms possess the underlying data, control access to audiences, determine the contractual conditions of participation, and frequently obtain broad licenses over user-generated content. Economic value generated through collective participation is therefore captured primarily at the platform layer. Grassroots architectures alter this allocation. Participants possess their own data, retain ownership of their digital speech acts, control dissemination through cryptographic infrastructure, and may independently determine how those assets are used. The result is not merely greater privacy or autonomy, but a redistribution of economic agency itself. 54 A. Grassroots Architectures as Property-Owning Networks The preceding discussion argued that digital speech acts increasingly function as productive assets within contemporary society. The economic implication of that conclusion is straightforward: ownership of digital speech acts determines who may capture the value those assets generate. Rawlsâs account of property-owning democracy was concerned with dispersing productive assets throughout society rather than concentrating them within a small number of institutions. Grassroots architectures may be understood as extending that principle into the digital economy. Rather than concentrating ownership of expressive and informational assets within platform firms, they distribute ownership among the participants who create, possess, and maintain those assets. This redistribution alters the economic structure of participation. Under platform architectures, users contribute content, attention, behavioral information, and social interaction to systems owned by others. Participation generates value, but ownership of the infrastructure and associated assets remains concentrated. Under grassroots architectures, people remain participants while simultane- ously occupying the position of owners. They possess the expressive assets through which value is generated and retain the legal rights necessary to determine how those assets may be used. The economic significance of this distinction extends beyond any particular revenue model. Ownership provides the legal foundation upon which future forms of exchange, licensing, coopera- tion, and commercialization may be built. Participants become economic actors rather than merely sources of platform-generated value. 1. Economic Autonomy and Creator Commercialization The combination of possession and copyright ownership creates forms of economic autonomy unavailable within intermediary-controlled systems. First, creators may license digital speech acts directly. Copyright ownership permits authors to authorize reproduction, distribution, adaptation, and other uses of their expressions. Grassroots architectures place those rights in the hands of participants rather than intermediary institutions. Creators may therefore negotiate directly with readers, viewers, publishers, applications, or other entities seeking access to expressive works. Second, creators may engage in direct sales or subscription arrangements. Rather than relying upon platform-mediated subscription or advertising systems, participants may condition access to content on payment, membership, contractual participation, or other arrangements determined by the creator. Whether implemented through traditional payment systems, digital currencies, or cooperative structures, the legal authority to establish such arrangements derives from ownership itself. Third, creators may organize collectively. Data cooperatives and related governance structures have been proposed as mechanisms through which individuals can exercise collective control over informational resources. 267 268 Copyright ownership of digital speech acts creates similar possibil- ities for collective licensing, rights management, syndication, and cooperative bargaining. Rather than assigning rights to platforms, creators may pool rights while retaining collective control over their use. Fourth, ownership preserves attribution and reputation. Economic value frequently arises not through direct payment but through recognition, audience development, professional opportunities, and social influence. Where attribution remains cryptographically bound to expression, participants retain the reputational value generated through their contributions rather than allowing that value to be absorbed into platform-controlled ecosystems. These possibilities do not depend upon any single business model. The central point is more modest but more fundamental: ownership creates the legal capacity for creators to determine how value generated by their expressions will be captured and distributed. 267 See SalomĂ© Viljoen, A Relational Theory of Data Governance, 131 Yale L.J. 573 (2021). 268 See also Thomas Hardjono et al., Toward a Design Philosophy for Interoperable Blockchain Systems, MIT Connection Science (2018). 55 2. Intermediary Rent Extraction and Alternative Market Structures The concentration of ownership within platform architectures has broader consequences for market organization. Platform firms derive economic power not solely from producing content themselves, but from controlling the infrastructure through which content circulates. This position enables platforms to capture value generated by billions of users whose participation makes the platform economically valuable. In this respect, platform ownership resembles other historical forms of concentrated ownership in which control over productive infrastructure permits extraction of economic rents from the activities of others. 269 Grassroots architectures reduce this intermediary position by relocating ownership and possession to participants themselves. Economic relationships need not pass through a central platform because participants retain both the assets and the infrastructure necessary for exchange. This does not eliminate intermediaries entirely. Applications, search services, reputation systems, discovery mechanisms, and other forms of coordination may continue to emerge. The distinction is that such intermediaries would compete for access to participant-owned assets rather than exercising ownership over those assets themselves. Economic power would therefore arise from providing services rather than controlling participation. The result is the possibility of alternative market structures in which value capture occurs closer to the individuals and communities generating value in the first place. The significance of intermediary rent extraction therefore extends beyond questions of market structure or business models. As Part I argued, platforms function as juridical infrastructures that allocate ownership, possession, and authority through the interaction of copyright, contract, and technical architecture. The rents extracted by platforms are not merely economic outcomes but legal and institutional consequences of those underlying arrangements. Because users lack possession of the infrastructure, surrender broad contractual rights, and participate through systems owned by others, value generated through their expressive activity becomes available for intermediary appropriation. Grassroots architectures alter this allocation. By reuniting ownership and possession in partici- pants and reducing reliance upon intermediary-controlled contractual regimes, they redistribute not only economic value but also the legal authority through which that value is captured. The question is therefore not simply who earns revenue from digital participation, but which juridical infrastructure governs the production and distribution of that revenue in the first place. 3. Limitations and Open Questions These economic possibilities remain subject to substantial practical limitations. Transaction costs remain significant. Direct licensing and compensation require payment infras- tructure, contracting mechanisms, and systems for discovery and enforcement. 270 Discovery presents additional challenges. Platforms currently provide recommendation systems, search functionality, and audience aggregation. Grassroots architectures must develop alternative mechanisms for connecting creators with audiences without recreating the concentration of power they seek to avoid. Collective-action and free-rider problems also persist. Open sharing may reduce incentives for payment-based systems, while restrictive licensing may reduce the benefits of network participation. Determining appropriate balances between openness and compensation remains an unresolved design question. Finally, network effects continue to favor established platforms. 271 Ownership alone does not guarantee adoption. Users frequently value convenience, familiarity, and existing social connections. Whether distributed ownership can overcome these advantages remains an empirical question. 269 See generally Brett Frischmann, Infrastructure: The Social Value of Shared Resources (2012). 270 See Joseph Bonneau et al., SoK: Research Perspectives and Challenges for Bitcoin and Cryptocurrencies, IEEE Symp. on Security & Privacy 104 (2015). 271 See Mark A. Lemley & David McGowan, Legal Implications of Network Economic Effects, 86 Cal. L. Rev. 479 (1998). 56 Copyright ownership of digital speech acts does not guarantee any particular economic outcome. It does, however, establish the legal preconditions necessary for participants to capture value generated through their own digital activity rather than ceding that value by default to intermediary institutions. XII. CONCLUSION This Article began with a puzzle. Contemporary platform firms derive extraordinary economic value from user-generated expression, interaction, and participation, yet ownership of the produc- tive assets generating that value remains concentrated within intermediary institutions rather than distributed among the individuals who create them. Digital expressions have become economi- cally productive resources, but existing platform architectures separate creation from ownership, participation from control, and value generation from value capture. The conventional response treats this as a regulatory problem requiring platform reform: trans- parency mandates, content moderation oversight, algorithmic accountability, and procedural safe- guards. Those interventions may constrain platform power, but they generally leave intact the underlying architecture through which platforms mediate communication and accumulate control. This Article has explored a different possibility: an architectural alternative in which people cre- ate, possess, own, and exchange digital expressions without surrendering authority to centralized intermediaries. The legal foundation for that alternative rests on existing copyright doctrine rather than regulatory innovation or law reform. Digital speech acts satisfy the traditional requirements of authorship, originality, and fixation in copyright law as reflected in Burrow-Giles Lithographic Co. v. Sarony, Feist Publications, Inc. v. Rural Telephone Service Co., and the Copyright Act. Part VIII demonstrated that copyright ownership persists through peer-to-peer dissemination and that cryptographic provenance enables attribution across forwarding chains. These capabilities generate both opportunities and challenges. Accountability becomes more direct, but questions concerning intermediary immunity, jurisdiction, liability, and enforcement become more complex. Grassroots architectures do not eliminate legal problems; they relocate them. Courts confronting digital speech acts for the first time may reject copyrightability on grounds of functionality, merger, minimal creativity, or broader policy concerns. Part IX engaged those objections directly. Part X argued that ownership of digital speech acts is justified not merely as a matter of labor- desert but as a matter of democratic structure. Political theorists have long recognized that ownership of productive assets affects the conditions under which citizens participate in society. Rawlsâs concern was not with any particular form of capital but with preventing domination through concentrated ownership. In digital societies, expressive and informational assets increasingly perform productive functions analogous to those once performed by land, factories, and machinery. Ownership of those assets therefore affects the distribution of economic power, political influence, and social participation. The significance of grassroots architectures lies in their reunification of ownership and possession. Participants retain copyright in their expressions while simultaneously maintaining possession of the devices and data through which those expressions are created and exchanged. This combination creates a form of digital independence unavailable within intermediary-controlled systems. Rather than participating as tenants within privately governed platforms, individuals participate as owners of the productive assets through which digital life increasingly occurs. The economic implications are equally significant. Ownership of digital speech acts does not merely allocate control over participation; it allocates the capacity to capture economic value gener- ated through participation. Grassroots architectures therefore create legal conditions for alternative forms of economic organization, including direct licensing, creator-controlled commercialization, cooperative ownership structures, collective bargaining arrangements, and attribution-based repu- tation economies. Whether such arrangements ultimately succeed remains an empirical question. What matters for present purposes is that distributed ownership creates possibilities unavailable when ownership is concentrated within platform institutions. 57 A. What This Establishes This Article establishes four propositions. First, digital speech acts are copyrightable under existing doctrine. Human authors make the expressive choices that generate digital speech acts; cryptographic systems merely execute those choices. Existing precedent therefore provides a substantial basis for recognizing authorship, originality, and fixation. Courts may disagree, but current doctrine does not foreclose the claim. Second, copyright ownership combined with device-based architecture is a prerequisite for digital sovereignty and economic agency. When individuals own their expressions, possess their data, and participate through systems they control, they gain legal and practical authority over assets that would otherwise be subject to intermediary control. Third, digital speech acts increasingly function as productive assets within contemporary society. The democratic question is therefore not merely whether users enjoy privacy or autonomy, but whether ownership of economically productive expressive assets remains broadly distributed or becomes concentrated within a small number of institutions. Fourth, the legal and theoretical foundations for grassroots architectures are sufficient, even if practical success remains uncertain. Copyright doctrine supports person ownership. Democratic theory justifies distributed control. Technical architectures capable of implementing these principles exist. Whether such systems achieve widespread adoption remains an empirical question rather than a legal one. B. Architecture, Ownership, and Power The central insight of this Article is that copyright law is largely neutral regarding architecture. The same legal system that facilitates platform dominance can also support person ownership. The critical variable is not copyright doctrine itself but the institutional arrangements through which copyright operates. Server-based architectures concentrate possession, control, and governance within intermediary institutions. Device-based architectures distribute possession and permit ownership to remain with participants. Copyright follows those architectural facts. When people create expressions, possess the underlying data, and retain legal rights in their works, ownership remains distributed. When plat- forms aggregate possession, govern access, and acquire extensive rights through contract, ownership becomes effectively centralized. The question is therefore not whether copyright inherently empowers platforms or users. It empowers whichever actors possess the practical and institutional capacity to control expressive assets. Architecture determines where that capacity resides. C. The Stakes The stakes extend beyond social networking. Platform-mediated digital life increasingly struc- tures communication, commerce, culture, association, and civic participation. As a result, questions of ownership increasingly become questions of democratic organization. The central democratic question is not simply who may speak online, but who owns the productive assets through which participation occurs. The concentration or distribution of expressive and informational assets determines the conditions under which citizens participate in economic, social, and political life. Conventional regulatory responses generally accept platforms as inevitable and seek to constrain their power through oversight and accountability mechanisms. Those reforms may be valuable. They do not address the deeper issue of dependency created by concentrated ownership. Grassroots architectures offer a different possibility. By combining distributed possession, person-owned copyright, and voluntary contractual governance, they create conditions under which participation need not depend upon intermediary permission. They replace relationships of depen- dency with relationships of ownership. Grassroots systems will encounter governance challenges, moderation difficulties, economic 58 uncertainties, technical limitations, and adoption barriers. Distributed architectures redistribute power rather than concentrate it, creating institutional conditions more consistent with democratic ideals of independence and self-government. D. Democratic Alternatives Are Possible This Articleâs ultimate claim is one of possibility rather than inevitability. Contemporary plat- forms are not the only viable architecture for digital social life. Existing copyright doctrine can support person ownership of digital speech acts. Democratic political theory provides reasons to value such ownership. Technical architectures exist that can implement these principles without relying upon centralized intermediaries. The question is no longer whether digital participation generates productive assets. Contemporary platforms have already demonstrated that it does. The question is who will own those assets. If ownership remains concentrated within intermediary institutions, digital society may continue to reproduce dependency, extraction, and centralized control. If ownership remains with the individ- uals who create, possess, and exchange digital expressions, different forms of governance, economic organization, and democratic participation become possible. Copyright law does not dictate which future emerges. It provides legal foundations upon which either may be built. 59