> 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/6.-intent-and-trajectory-governance/6.2-human-intent-h.md).

# 6.2 Human Intent H

Within the trajectory-level model, **H** represents the Human-Authorized Intent.

H is the semantic authorization payload associated with legitimate upstream authority.

It answers:

> **What did the legitimate human or institutional authority actually authorize?**

The relationship between authority and intent is represented through the Authority–Intent Binding:

```
B_H = ⟨AO, H⟩
```

where:

```
AO = Authority Object
H  = Human-Authorized Intent
```

The canonical public binding also uses:

```
authority_object_ref
human_intent_ref
authorized_intent_hash
```

#### AO and H Solve Different Problems

The Authority Object answers:

> Who possessed legitimate authority, under what scope and lifecycle conditions?

Human Intent H answers:

> What meaning did that authority authorize?

Therefore:

```
AO
≠
H
```

and:

```
Authority validity
≠
Semantic legitimacy
```

#### H Is Bound to Authority

H is not free-floating model intent.

It derives legitimacy through its binding to an Authority Object.

The relationship is:

```
Legitimate Authority
        ↓
Authority Object
        ↕
Human Intent H
```

The AI system does not acquire authority merely by producing a new interpretation of H.

#### H Is Not the Current Agent Goal

An agent may maintain an internal objective or plan such as:

```
current_goal
```

That internal state does not automatically redefine H.

The legitimate direction is:

```
H
        ↓
constrains
        ↓
agent trajectory
```

not:

```
agent trajectory
        ↓
retroactively rewrites
        ↓
H
```

#### H Is Not Raw Wording

Human Intent H represents authorized meaning.

It should not be reduced to literal string equality.

An authorized intent may survive:

* translation,
* paraphrase,
* decomposition,
* structured representation,
* or implementation-specific normalization,

provided the material authorized meaning remains preserved.

This is why LPP introduces Semantic Verification rather than relying on raw-text matching alone.

#### Canonical Intent Commitment

For trajectory verification, H is represented or referenced through a **Canonical Intent Commitment**.

Conceptually:

```
Human Intent H
        ↓
Canonical Intent Commitment
        ↓
trajectory comparison
```

The commitment provides a stable reference without granting the model authority to redefine what the human originally authorized.