Substrate · Overview

The Encoding Stack

VCP requires substrate to be machine-parseable. The substrate is a three-layer encoding stack — lexical, structural, relational.

Version
v0.1
Status
Active
Verified
12 May 2026
Cite
valuecreationprotocol.com/encoding-stack
Three layers: Lexicon (lexical) → VCP-Lang (structural) → Value Graph (relational)
Each layer makes a different aspect of VCP machine-executable.

Why an encoding stack

Foundation models default to industrial-age vocabulary because that vocabulary dominates training data. Prompt engineering can shift outputs at the surface, but the underlying defaults reassert under load — under context pressure, ambiguity, or when the model has to reason about novel situations.

The encoding stack gives VCP a substrate that does not depend on prompt-level overrides. Each layer makes a different aspect of VCP machine-executable.

The three layers

How the layers work together

A VCP-aligned system declares its values, context, and trust commitments using VCP-Lang, references terms whose meaning is fixed by the Lexicon, and operates over relationships modeled in the Value Graph. An AI system reading those declarations can verify alignment, surface drift, and generate outputs that respect the protocol without re-deriving the methodology from natural-language description.

Status

All three layers are at v0.1 — defined, scoped, and entering implementation. The Lexicon has the longest history (descended from a multi-year language guide). VCP-Lang is the newest layer. The Value Graph is being formalized from existing relational frameworks.

The encoding stack is not VCP. It is the substrate that makes VCP executable. VCP can be referenced without invoking the encoding stack; the encoding stack should not be referenced without invoking VCP, because the stack exists to serve the protocol.

Protocol home

VCP is originated and canonically implemented by Value-First Team. Anyone may read, cite, and operate the protocol independently of firm engagement.