Plain Language Guide

When the Architecture of Law Predicts Its Outcomes

A layperson's guide to the SOoL structural jurimetrics study — no legal or technical background required

0.9848

AUC — Structural Consistency (masked, internal)

Stable under adversarial audit — 18% of high-CD cases flagged as structurally contested; figures shift <0.003 AUC under maximum correction scenario.

14.0×

CD Ratio (Denied / Granted)

4,931

Federal Appellate Cases

0.9977

AUC — Temporal Hold-out

"An architect asks whether the building is beautiful. A structural engineer asks whether it will stand. SOoL is the structural engineering approach applied to law."

The Core Idea: Structural Soundness vs. Doctrinal Beauty

Standard legal analysis asks: does the rule apply to these facts? SOoL asks something more fundamental: is the architecture of this legal claim structurally sound? A claim can have excellent facts and strong doctrine and still fail — if the structure holding it together is broken.

The Eight-Step Structural Checklist (The MLC)

MLC stands for Minimum Legal Chain — the minimal sequence of eight conditions that must all be satisfied for a legal claim to be structurally complete. Think of it as eight load-bearing walls. If any wall is missing or cracked, the structure is at risk.

N1 — Source of Authority

Is there a legitimate legal source — a statute, constitutional provision, or binding precedent — that authorizes this rule?

N2 — Norm

Is there a determinate, intelligible rule that specifies what conduct is required, permitted, or prohibited?

N3 — Actor in Role

Is there an identifiable person bearing a specific legal role — employee, officer, regulated entity — to whom the rule applies?

N4 — Triggering Facts

Are the facts that activate the rule actually present and accurately established? Fabricated or suppressed facts fail here.

N5 — Legal Act or Omission

Did the relevant conduct actually occur? Courts often disagree about whether particular behavior counts as the legally relevant act.

N6 — Target

Is there an identifiable party against whom the obligation runs, and do they have legal standing to be in the case?

N7 — Legal Effect ★ PIVOTAL

Does the right, duty, or legal consequence actually attach? This is where most claims ultimately succeed or fail — the recognition gateway.

N8 — Remedy

Is there an available remedy if the obligation is breached? A right without a remedy is structurally incomplete. Always fails when N7 fails.

Contradiction Debt (CD): The Structural Stress Score

CD stands for Contradiction Debt — a single number that measures how structurally stressed a legal claim is. It is calculated by identifying which structural failure modes are active in the case and adding up their assigned weights. A CD of zero means the legal architecture is sound. The higher the CD, the more structural stress.

Crucially, CD is not circular — it does not simply re-encode the outcome. When we tested a model using only the upstream nodes (N1–N6), ignoring the recognition and repair nodes entirely, it predicted outcomes at AUC = 0.5518 — barely better than a coin flip. The near-perfect prediction from the full CD score is emergent from the whole chain structure, not from reading the conclusion and working backward. An independent adversarial audit — in which a second pass explicitly argued the strongest structural case for the opposite outcome for every high-CD annotation — flagged 17–18% of cases as structurally contestable. Even under a maximum-correction scenario where every flagged case were re-annotated, the AUC would shift by less than 0.003. The finding is robust to adversarial pressure.

CD STABILITY SCALE

0.00 — Stable (clean chain) 0.15 — Stressed 0.30 — Unstable 0.45+ — Collapse risk

0.013

Mean CD — Protection Granted (n=1,027)

0.257

Mean CD — Protection Denied (n=3,265)

Four Independent Validation Tests

We tested the predictive validity of CD four different ways to make sure the result was real. Every test confirmed it.

TEMPORAL HOLD-OUT TEST

AUC = 0.9977

Trained on pre-2020 cases only. Tested on 2020–2024 cases the model had never seen. The structural signal does not decay as judges change — mean CD for losing cases: 0.258 before 2020, 0.259 after.

LEAVE-ONE-DOMAIN-OUT TEST

All 8 domains ≥ 0.99

Trained on seven doctrinal areas, tested on the eighth. A model trained on Contract and Criminal Procedure cases predicts First Amendment outcomes at AUC = 0.9996. The structural signal crosses doctrinal boundaries.

UPSTREAM-ONLY TEST (KEY RESULT)

AUC = 0.5518 ≈ Chance

Using only the first six nodes — excluding the recognition and remedy nodes — the model barely beats a coin flip. This directly refutes circularity: the annotation is not secretly encoding the outcome.

RANDOM HOLD-OUT TEST

AUC = 0.9848

Masked re-annotation of full corpus (n=4,934, disposition language stripped). Primary non-circular validation figure. Chain_outcome agreement with unmasked: 96.1%.

Where Law Fails: Three Structural Clusters

Different areas of law fail at different structural nodes. This is not random — it reflects the formal architecture of each doctrinal area. The data reveal three clusters:

N2 CLUSTER — NORM INDETERMINACY

D7 Contract · D8 Family Law

The governing rule itself is contested or open-textured. Which body of law applies? What does "best interests of the child" actually require? The architecture fails because the norm cannot be specified.

N5 CLUSTER — CONDUCT CLASSIFICATION

D1 First Amend · D2 Employment · D6 §1983

The central question is whether the specific act performed counts as the legally relevant act. Does this speech fall within protected expression or official-duty conduct? (Garcetti v. Ceballos is the paradigm case.)

N6 CLUSTER — TARGET RECOGNITION

D3 Admin Law · D5 Immigration · D4 Criminal

The central question is whether the party before the court can assert the right at all. Does this company have standing to challenge the agency? Is this immigrant recognized as a rights-bearing subject or only as an object of enforcement?

What This Means for Lawyers, Theorists, and Students

⚖ For Practitioners

The MLC audit identifies structural weak points before filing. A case with RC3 at N3 will propagate stress to N7 regardless of factual strength. Knowing where the chain is weak lets you reframe, fix, or factor the structural risk into settlement valuation.

📚 For Theorists

The temporal hold-out result (AUC = 0.9977 across a 30-year span, stable CD across court compositions) directly challenges legal realism: outcomes are not primarily determined by who the judges are. The structural architecture of law predicts outcomes more reliably than judicial ideology.

🎓 For Students

IRAC tells you what the law says. The MLC tells you whether the structure holding the argument together is sound. You will spend your legal career applying rules — understanding the structural architecture beneath those rules makes you a better analyst of why some arguments win and others lose.

Why the Weights Are What They Are

Every structural failure mode carries a weight that determines how much it contributes to the Contradiction Debt score. These weights were assigned from first principles before any data was collected — they were not calibrated to predict outcomes. They reflect three theoretical criteria: how fundamentally the failure violates the BFO-typed constraint it targets; how far down the chain the failure occurs and how irreversible it is; and how difficult the structural problem is to repair without external intervention.

The weights range from 0.07 to 0.18 — a ratio of 2.57:1 between the lowest and highest. This spread is deliberate: narrow enough that all thirteen types remain informative, wide enough to reflect genuine theoretical differences in severity.

CODE WEIGHT NAME & PLAIN-LANGUAGE RATIONALE NODE

RCL 0.18 Recognition Collapse — the most severe violation N6→N8

The highest weight because RCL destroys the symmetry of the legal relationship itself. The agent loses the capacity to hold rights while retaining the capacity to bear obligations — the same person is simultaneously unable to assert protections and can still be held to duties. This violates the fundamental Hohfeldian principle that rights and correlative duties must exist together. No other failure type destroys this ontological symmetry so completely. It also carries the heaviest repair burden: restoring someone's legal subjecthood requires institutional reconstitution, not just doctrinal clarification. Paradigm: an immigrant in removal proceedings who is regulated as an object of enforcement but cannot assert rights against removal.

CF 0.15 Conferral Failure — the chain has no foundation N1→N3

The second highest because it fails at the very first link of the chain and does so irreversibly within the proceeding. If the authority source failed to properly vest the actor in the required role, there is no legitimate actor — the entire subsequent structure has no bearer. You cannot fix this by clarifying the norm or reassessing the facts. The chain simply has no foundation. In patent law, this fires when the patent is found invalid: the USPTO failed to properly confer the right, so the authority source at Node 1 never properly existed. High repair burden, earliest possible chain position, complete structural void.

RC3 0.14 Role Contradiction — incompatible obligations for the same act N3

The same agent bears two roles whose obligations are genuinely incompatible for the same act — not merely difficult to reconcile, but structurally irreconcilable. The structural problem is not vagueness but genuine incompatibility: no amount of normative clarification resolves it. The only structural resolution is role recharacterization, which typically requires a new legal framework rather than better application of the existing one. Garcetti v. Ceballos is the paradigm: Richard Ceballos simultaneously holds the citizen-speaker role (whose speech is protected) and the official-duty public employee role (whose work product is not) for the exact same memo. No interpretation of the Pickering/Garcetti framework can honor both roles for the same act. This is why RC3 appears in 19.5% of First Amendment corpus cases — the structural problem is endemic to the doctrinal architecture.

CC 0.14 Correlativity Contradiction — right acknowledged, duty refused N7

Matches RC3 because both sever a Hohfeldian correlation, but CC does it at Node 7: the court acknowledges the right exists and was violated, then refuses to enforce the correlative duty through immunity or structural bar. The court is effectively saying: your right was violated, and the person who violated it faces no obligation to answer for it. This is ontologically impossible in a coherent legal system — a right without a correlative duty is not a right in any meaningful sense. The qualified immunity doctrine in civil rights cases is the paradigm: "Assuming a constitutional violation occurred, the right was not clearly established, so the officer is immune." Right acknowledged; duty refused. The 0.14 weight reflects that this is a recognized structural contradiction built directly into doctrine — not an accident but a deliberate structural choice with serious ontological consequences.

RPF 0.13 Repair Failure — remedy structurally blocked N8

High weight but below CF and RCL because RPF almost always cascades from Recognition Failure rather than firing independently. When Node 7 fails (right not recognized), Node 8 necessarily fails too — the claimant has no recognized right to repair. The corpus confirms this at a phi correlation of 0.933. The weight is set high enough to reflect that blocked repair is genuinely serious, while acknowledging the cascade nature of most RPF findings. The important exceptions — where RPF fires independently — are the most structurally interesting cases: habeas corpus (the entire proceeding challenges prior repair adequacy), FRAND patent licensing (the remedy pathway is explicitly blocked by licensing obligations), and some qualified immunity cases where relief is theoretically available but procedurally inaccessible.

AI 0.12 Authority Inflation — personal will displaces the norm N1→N2

Higher than NI and JC because AI is not a problem of vagueness or competing norms — it is a displacement of the institutional norm by personal will. When a decision-maker's preferences substitute for the governing rule, the chain loses its institutional grounding at the N1→N2 link. The authority source continues to exist formally, but the norm it issues is no longer traceable to that source — it is traceable to the individual's own disposition. This severs the relationship that makes norms institutional rather than personal. Post-Loper Bright, AI is expected to become more common in administrative and securities cases as agencies attempt to extend their authority through enforcement action rather than rulemaking — using prosecutorial discretion to create new norms without going through the statutory authority process.

RF 0.11 Recognition Failure — court denies the right (most common CT) N6/N7

Lower than RCL and RPF despite appearing in roughly 65% of corpus cases. This reflects a core theoretical principle: frequency does not imply severity. RF fires when the court denies the claimant's asserted right — the legal effect fails to be recognized. This is structurally serious and consequential, but no Hohfeldian correlation is violated: there is no acknowledged right without a correlative duty, because the right is simply denied. The court is operating coherently within the system — it has evaluated the claim and found it insufficient. The relatively modest weight reflects that RF is the normal operation of a legal system that evaluates claims and sometimes rejects them, rather than a fundamental ontological rupture.

FM 0.11 Fact Manipulation — the factual foundation is corrupted N4

Matches RF because both represent serious but structurally bounded failures. FM fires when the triggering facts are fabricated, suppressed, or made indeterminate — the factual foundation of the entire chain is corrupted. The paradigm is a Brady violation: the prosecution withholds evidence that would help the defendant, making the triggering facts structurally indeterminate for the defense. FM is weighted at 0.11 rather than higher because it is in principle remediable — the structural problem is an epistemic failure about facts that, if correctly established, might support a coherent chain. This is qualitatively different from RCL (which requires reconstituting legal subjecthood) or CF (which voids the authority chain). A retrial with correct facts can repair FM. Nothing comparable repairs RCL.

JC 0.10 Jurisdictional Contradiction — two valid norms, incompatible claims N2→N6

Moderate weight because both competing norms are individually legitimate — the problem is their coexistence, not their individual validity. Two authority sources each issue a norm that claims to govern the same conduct, and the norms pull in incompatible directions. This is structurally serious but in principle resolvable through conflict-of-laws analysis, preemption doctrine, or legislative clarification. Neither norm is invalid; the system simply has not yet resolved which one governs. Federal/state conflicts and civil/criminal regime overlaps are the paradigm cases. JC is more remediable than RC3 (where a single actor's roles are irreconcilable) because the resolution mechanism — deciding which norm prevails — is well-established in legal doctrine even if applying it is difficult.

SE 0.10 Self-Undermining Effect — the right destroys the system that recognizes it N7

Matches JC because both involve systemic rather than case-level structural problems. SE fires when the legal effect at Node 7, if fully realized, would destroy the authority source at Node 1 that generates it — a self-referential collapse. The realized right undermines the institutional conditions for its own recognition. This is ontologically precise: the effect's realization destroys the continuants that ground its own authority. SE is weighted moderately because it is rare, highly context-dependent, and philosophically complex — it requires a case where recognizing the right would destabilize the system's capacity to recognize any rights. Cases involving claims against the structural prerequisites of judicial authority itself are the paradigm.

PC 0.09 Procedural Contradiction — the norm blocks its own fulfillment N2→N5

The act required by the norm cannot be performed through the norm's own procedure — a self-blocking realization pathway. Lower weight than NI (which questions what the norm is) and JC (which involves competing authority sources) because PC is a procedural rather than substantive failure. The substantive norm may be perfectly sound; the procedural mechanism that implements it is defective. This is remediable through procedural reform without touching the substantive norm itself. An example: a statute that requires claimants to exhaust administrative remedies before suit, where the administrative remedy process is itself unavailable or takes longer than the statute of limitations allows. The norm is valid; the procedure defeats it.

TC 0.08 Temporal Contradiction — norm and facts misaligned in time N2↔N4

Second-lowest weight because temporal misalignment is the most readily remediable structural failure. TC fires when the norm and triggering facts are misaligned in time — either the norm was not yet in force when the facts arose (retroactivity) or the facts predated the norm's existence (ex post facto). The structural fix is obvious and requires no fundamental ontological reconstitution: the norm simply should not apply to these facts. TC requires temporal boundary clarification, not institutional reconstitution. This is qualitatively different from RCL or CF. A statute that purports to apply retroactively has a structural TC problem, but the remedy — holding it inapplicable — is clean and leaves the rest of the legal system undisturbed.

NI 0.07 Norm Indeterminacy — the governing rule cannot be specified (lowest) N2

The lowest weight for two reasons. First, courts routinely resolve NI in the very opinion that identifies it — they specify the indeterminate norm and apply it, turning an NI into a closed node within the same proceeding. The Garcetti majority reservation (declining to decide academic speech) is a rare case of genuine persistent NI; most NI findings are resolved by the time the opinion issues. Second, NI is often not a failure of the legal system but a recognition that the system has reached the frontier of its current normative development. A norm that is indeterminate today may become determinate through legislative or judicial specification tomorrow. This is qualitatively different from RCL, which strips rights that the system cannot simply declare back into existence, or CF, which voids an authority chain that must be rebuilt from the institutional ground up. NI is structurally significant — an indeterminate norm cannot activate the chain — but it is the most remediable of the thirteen failures.

CD WEIGHT SPECTRUM — CLICK ANY ROW ABOVE TO EXPAND RATIONALE

NI 0.07

TC 0.08

PC 0.09

JC 0.10

SE 0.10

RF 0.11

FM 0.11

AI 0.12

RPF 0.13

RC3 0.14

CC 0.14

CF 0.15

RCL 0.18

Weights range from 0.07 (NI — most remediable) to 0.18 (RCL — most severe). The 2.57:1 spread keeps all thirteen types informative while reflecting genuine theoretical differences. These weights were fixed before any data was collected — the empirical 14.0:1 CD/outcome ratio is a validation of the theoretical framework, not a product of weight optimization.

Quick Glossary

AUC Area Under the Curve — predictive accuracy measure. 0.5 = coin flip; 1.0 = perfect.

BFO Basic Formal Ontology — a philosophical framework for classifying what kinds of things exist.

CD Contradiction Debt — a numerical score measuring structural stress in a legal chain.

CT Contradiction Type — one of thirteen specific structural failure modes indexed to MLC links.

IRAC Issue, Rule, Application, Conclusion — standard legal analysis framework taught in law school.

MLC Minimum Legal Chain — the eight-node structural checklist for legal claim completeness.

RF Recognition Failure — court denies the claimed right exists or applies (most common CT).

RPF Repair Failure — remedy is structurally blocked. Always cascades from Recognition Failure.

RC3 Role Contradiction — same person bears two roles with incompatible obligations. Garcetti paradigm.

CC Correlativity Contradiction — right acknowledged, correlative duty refused (qualified immunity).

NI Norm Indeterminacy — court cannot specify what rule governs (scope uncertainty, not application).

SOoL Structural Ontology of the Law — the framework (Koepsell, Palgrave Macmillan, forthcoming).

RCL Recognition Collapse — subject's capacity to bear rights stripped while obligations continue.

FM Fact Manipulation — triggering facts fabricated or suppressed. Brady violation is the paradigm.

A Structural Ontology of the Law · David R. Koepsell · SEAL Lab, Texas A&M University · Palgrave Macmillan, forthcoming
[email protected] · seal.tamu.edu/legal-kernel ·

Structural Signature Query

Cross-Domain Family Resemblance

CD Comparative Query

MLC Reference

When the Architecture of Law Predicts Its Outcomes

The Core Idea: Structural Soundness vs. Doctrinal Beauty

The Eight-Step Structural Checklist (The MLC)

Contradiction Debt (CD): The Structural Stress Score

Four Independent Validation Tests

Where Law Fails: Three Structural Clusters

What This Means for Lawyers, Theorists, and Students

Why the Weights Are What They Are

Quick Glossary

Structural Ontology of Criminal Law

Structural Reasoner

Ontology Browser

Case Analyzer

Hypothesis Laboratory