> For the complete documentation index, see [llms.txt](https://docs.lpp-minduniverse.org/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.lpp-minduniverse.org/lingua-pactum-protocol-lpp-documentation/10.-research-and-publications/10.5-paper-4-authority-objects.md).

# 10.5 Paper 4 — Authority Objects

#### Publication

**Authority Objects: Formal Specification for AI-Triggered Action Admissibility**

Author: Jason Liao\
Publication: Zenodo\
Version: v0.2\
Publication Date: July 2026\
DOI:

```
10.5281/zenodo.21468687
```

#### Research Question

> **What machine-verifiable object carries legitimate authority before execution?**

The first three papers depend on the existence of valid authority.

Paper 4 asks what that authority actually looks like as a governance object.

Its key observation is:

> **Without Authority Objects, the Admission Kernel would be a gate without a verifiable legitimacy object.**

#### Core Definition

An Authority Object is defined as:

> **A structured, machine-verifiable legitimacy carrier for AI-triggered execution.**

It possesses five essential properties.

**Externally Grounded**

It originates outside the model's self-generated reasoning.

**Scoped**

It binds authority to relevant action, subject, domain, jurisdiction, or execution boundaries.

**Lifecycle-Bound**

It may be:

* minted,
* degraded,
* expired,
* revoked,
* or invalidated.

**Verifiable**

It carries or references appropriate validation material such as:

* signatures,
* policy bindings,
* consensus evidence,
* or other verification state.

**Admission-Bearing**

It can be evaluated by the Admission Kernel before execution.

#### Formal Object

The paper defines a minimal Authority Object of the form:

```
AO = ⟨
    id,
    subj,
    issuer,
    scope,
    jurisdiction,
    class,
    sig_req,
    sig_state,
    window,
    decay,
    revocation,
    transfer,
    policy,
    consensus,
    status
⟩
```

The canonical lifecycle domain includes:

```
UNMINTED
MINTED
INVALID
EXPIRED
REVOKED
```

#### Authority vs Access vs Permission

Paper 4 sharply separates:

```
Access
Permission
Authority
Admissibility
```

The distinction becomes:

```
Access:
Can the system reach the resource?

Permission:
Does policy permit the operation?

Authority:
Has a legitimate issuer granted a scoped,
time-bound, revocable basis for the action?

Admissibility:
Does this specific action remain qualified
to enter execution governance?
```

The paper therefore reinforces:

> **Access is not authority.**

> **Permission is not admissibility.**

> **Logging is not legitimacy.**

> **Model explanation is not authorization.**

#### Authority Object vs Execution Permit

Paper 4 formalizes a crucial two-object distinction:

```
Authority Object
=
Why legitimate permission may exist upstream
```

```
Execution Permit
=
What constrained execution may occur downstream
```

The Permit does not create legitimacy.

It derives from a successful admission decision that relied on a valid Authority Object.

#### T0–T4

Paper 4 establishes the general LPP consequence and irreversibility classes:

```
T0
No material external effect

T1
Fully reversible

T2
High recovery cost / partially reversible

T3
Legally or operationally irreversible

T4
Sovereign / asset-critical / systemic-risk
```

These are cross-domain LPP classes.

Domain-specific models such as LPP-FIN may map them into domain-specific equivalents.

#### Domain Examples

The paper demonstrates how the Authority Object concept may be instantiated across domains including:

* finance,
* cyber operations,
* healthcare,
* robotics / physical systems,
* and enterprise automation.

This shows that the legitimacy object is general even when domain fields differ.

#### Bounded Temporal Degradation

Paper 4 also considers authority as time-sensitive rather than permanently valid.

Temporal degradation is treated as bounded governance behavior rather than as an unlimited model-generated score capable of replacing authority.

#### Forward Boundary to Paper 5

Paper 4 explicitly leaves several adjacent problems outside its scope, including full definitions of:

* SVNN semantic verification,
* latent intent mismatch,
* Admission Artifact verification,
* and related trajectory semantics.

In particular, SVNN appears only as a forward-referenced verification or consensus evidence field.

This creates the direct research dependency into Paper 5.

#### Research Role

Paper 4 establishes:

```
The Machine-Verifiable Legitimacy Layer
```

It answers:

> **What carries authority into the Admission Kernel?**