What is Lost When Legal Grammar Runs Without Judgment? | HackerNoon

1. What happens when the law speaks without a speaker?

   In regulatory, clinical, and financial domains, language increasingly operates without referent and without subject. Documents that once required authorship and deliberation are now generated, compiled, or executed by systems that simulate neutrality. This paper asks: What is lost when legal grammar runs without judgment? The core claim: The syntactic execution of legal norms permits the elimination of the ethical trace, erasing authorship and displacing accountability from the institutional text.

2. Key Concepts Executable Sovereign:

   The non-human instance that decides by running reglas compiladas (compiled rules), not by interpreting meaning. Regla Compilada: A type-0 formal production (cf. Chomsky 1965), now used to encode legal logic into executable language. Ethical Trace: Any linguistic residue that signals who is responsible, why a decision is made, and on what grounds.

3. Legal Displacement by Syntax

   Automated legal drafting, especially with LLMs, introduces four structural shifts that suppress judgment: Passive displacement: “It is required that…” instead of “The agency requires…” Nominalization masking: Turning decide into decision, comply into compliance - verbs lose time, actor, and force. Procedural vagueness: Conditions and obligations embedded in modals (“may,” “should,” “as needed”) without enforcement source. Referent opacity: Citations point to norms or frameworks, but not to a human or institutional agent who can be held liable.

In sum, the law becomes an interface. Authority is no longer issued - it is compiled.

4. Why This Matters (Legally)

   A. Responsibility is non-delegable Legal systems rely on attribution: who drafted, who approved, who enforces. A regulatory clause without traceable agency breaks the chain of liability. This has direct consequences for judicial review, institutional oversight, and democratic legitimacy. B. Interpretability is not neutrality By removing judgment from syntax, LLM-drafted legal text becomes syntactically valid but normatively vacuous. Courts, auditors, and stakeholders cannot trace intentions or objections embedded in the draft. C. Non-referential compliance breaks enforceability If no actor is named, and no discretionary threshold is visible, then legal compliance becomes a formal gesture, not a binding act. The rule is “executed,” not “applied.” D. Text-as-code is not a defense Institutions may claim that a document “was generated by the system,” but in legal epistemology, execution without authorship is inadmissible as a source of normativity. Traceability is not optional - it is constitutive.

5. Field Cases (Legal-First Perspective)

   DomainExample SentenceProblemRegulatory”All vendors must comply…”No named issuer or statute basePrivacy”Consent is assumed if the form is signed”Passive evasion of intentContracts”Payment shall be processed on completion”No agent, no procedural triggerCrypto TOS”Utility may be granted as adoption grows”Vagueness plus agency removalClinical”Patient was cleared for discharge”Legal exposure without subject

6. A Structural Test for Legal Drafting

   ✅ Does each binding clause name an accountable actor?

   ✅ Are obligations expressed in active voice with enforceable triggers?

   ✅ Is there a visible chain of authority from clause to approver?

   ✅ Are modals (“must,” “should,” “may”) governed by explicit rules?

   ✅ Can a reader audit authorship without external documentation? If the answer is no, then the syntax may be valid, but the legality is hollow.

7. The Paper: For Legal Scholars, Builders, and Reviewers

   “Grammar Without Judgment: Eliminability of Ethical Trace in Syntactic Execution”  provides: A formal model of regla compilada as syntactic authority A typology of ethical erasure in legal-technical texts Structural diagnostics to prevent institutional outsourcing of responsibility Theoretical grounding from Chomsky (1965), Montague (1974), and Startari (2025) Compatibility with EU AI Act Articles 28–30 and traceability obligations

📄 https://doi.org/10.5281/zenodo.15783365  📄 https://papers.ssrn.com/abstract=5334549

8. Call to Action

   If your team, organization, or institution uses LLMs or compiles rules into executable policy, audit the output now. Make the subject visible. Make the obligation enforceable.  If the syntax runs without a judge, the law has no voice. Tag this paper in your next compliance review.  Use the traceability test above.  And if you rewrite a clause that hides responsibility - publish it.

Author

Agustin V. Startari  Linguistic theorist and researcher in historical studies. Author of Executable Power, The Grammar of Objectivity, and Algorithmic Obedience.

ORCID: 0000–0002–5457–0541

ResearcherID: K-5792–2016

SSRN Author Page: https://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=7639915

Website: www.agustinvstartari.com

Ethos

I do not use artificial intelligence to write what I don’t know. I use it to challenge what I do. I write to reclaim the voice in an age of automated neutrality. My work is not outsourced. It is authored.