# Why meaning needs a standard

*Foundations · lesson 1 of 5 · ~10 min*

## What you will learn

Why copying data between systems quietly destroys meaning, and what "federated semantics" changes.

## The problem: integration by copying

Every organization runs dozens of systems, and every system holds its own version of the same concepts: a *customer* in the CRM, a *counterparty* in accounting, a *user* in the product database. Integration, as practiced today, means copying records between them and translating fields in flight.

Three things go wrong, always:

1. **Meaning is lost in translation.** A field called `status` means something different in every system, and the mapping lives in an ETL script nobody reads.
2. **Copies drift.** The moment data lands in a second system, there are two versions of the truth, and no declared rule about which one wins.
3. **Context evaporates.** The copy does not carry *why* the value is what it is, who asserted it, when, or under what assumptions.

Scale this up: companies integrating with companies, governments with hospitals, AI agents with all of the above. Every boundary crossing repeats the same three failures. The industry's answer has been bigger central schemas: one warehouse, one master data model, one ontology to rule them all. Central schemas fail socially before they fail technically: nobody accepts someone else's definition of their own reality.

## The Vercy stance: meaning stays home

The Meta-Universe standard starts from the opposite premise:

> Every world of meaning is sovereign. Meaning stays home; understanding travels.

Concretely:

- Each organization (or product, or agent, or state) is a **Universe**: it models its own reality, in its own vocabulary, under its own constitution.
- Nobody imports anyone else's rows. Instead, universes exchange **Projections**: views of knowledge, generated by the owner, shaped by explicit **Contracts**.
- Where two universes talk about the same thing (this person, this invoice), they negotiate an **Identity binding**, not a merge.
- Every change is an **Event** on an append-only timeline, so any past state of knowledge can be reconstructed and every assertion has provenance.

This is federation instead of integration: no apex authority, no silent loss, disagreement preserved rather than overwritten.

## Why now

Two forces make this urgent. First, AI agents are becoming first-class producers and consumers of knowledge: an agent that ingests copied, context-free data confidently automates mistakes. Second, the cost of modelling has collapsed: what needed a data-architecture team now needs a well-instructed agent and a Git repository. The bottleneck is no longer writing models; it is keeping their meaning trustworthy across boundaries. That is exactly the part the standard fixes.

## A taste of the real thing

The DevTeam.Games platform model (one of the production models you will meet throughout these courses) holds a frozen export of a retired task tracker, live clones of source code mastered in GitLab, and its own authored knowledge, side by side, with a machine-readable register declaring who is the master of each. Nothing in it can silently become "a second truth". That register is one page of YAML. The discipline it encodes is this whole standard in miniature.

## Key takeaways

- Copying data across boundaries loses meaning, creates drift and strips context.
- Central schemas fail because sovereignty is non-negotiable.
- Vercy federates sovereign universes: projections instead of copies, contracts instead of hope, events instead of amnesia.

## Go deeper

- [Vision and Manifesto](/spec/#00-foundation/Vision.md)
- [Principles](/spec/#00-foundation/Principles.md)
- [Overview: the whole standard, connected](/spec/#00-foundation/Overview.md)

Next: [The hierarchy of meaning](02-hierarchy.md)