# VITA regenerative economics V2

**Status:** Current owner-settled structure with experimental calibration

**Date:** 2026-09-03

**Production target:** Base L2. Sepolia is development and rehearsal only.

This paper succeeds the historical [`VITA_TOKEN_SPEC.md`](VITA_TOKEN_SPEC.md). It closes the
structural gap exposed by the live development journey: a participant can hold five VITA while the
deployed development vouch minimum is ten, and a one-time seed cannot sustain repeated useful
identity work. VITA therefore needs the same regenerative shape as Daily Knomee while remaining a
separate scarce capacity with its own calibration.

Nothing in this paper authorizes a contract deployment or broadcast. The existing Sepolia contract
is historical development infrastructure and does not define Base production economics.

## Owner-settled invariants

1. **VITA is renewable identity-evidence capacity.** Eligible participants receive bounded future
   VITA replenishment. An enrollment-only grant is not the complete product loop.
2. **A VITA stake is immutable history.** The participant, subject, evidence commitment, original
   amount, creation order, identifier, and provenance remain replayable. Later evidence derives an
   outcome; it never erases or silently rewrites the stake.
3. **Dividends are future VITA issuance.** Later independent evidence that supports a VITA stake can
   improve its author's future VITA replenishment rate or capacity. It is not a transfer from the
   later attester and is not conserved against later stake.
4. **Contradiction retains real downside.** Later independent contradictory evidence can produce a
   negative reputation consequence and slash locked VITA according to the selected policy.
5. **Registered oracle evidence has extra signed weight in both directions.** Compared with an
   otherwise equivalent ordinary independent attestation, aligned registered-oracle evidence has
   more promotional effect and contradictory registered-oracle evidence has more slashing effect.
   The advantage is symmetric in structure. Its magnitude is not yet approved.
6. **Oracle evidence is not oracle adjudication.** An oracle appends a signed, provenance-bearing
   attestation. A frozen deterministic projector replays all admitted history through a cutoff and
   derives the result. No oracle unilaterally resolves, deletes, refunds, or rewrites a claim.
7. **Self-confirmation does not pay.** Evidence from the stake owner, duplicate evidence from the
   same independent source, and self-controlled oracle activity contribute no positive dividend.
8. **Stake size remains stake size.** Participant reputation does not multiply each VITA already
   staked. Reputation changes future VITA opportunity. Oracle evidence weight is a property of the
   admitted evidence class, not a multiplier on the voucher's tokens.
9. **Every economic choice is visible.** Recharge, wallet cap, stake lifetime, evidence weights,
   thresholds, slashing, and future-issuance functions must be identified in a parameter manifest.
   There are no undocumented defaults.

## Working state model

For VITA stake `i`, preserve an immutable record:

`S_i = (u_i, s_i, e_i, W_i, r_i, t_i, provenance_i)`

where `u_i` is the voucher, `s_i` is the canonical subject, `e_i` is the exact evidence commitment,
`W_i` is the original VITA amount, and `r_i` and `t_i` preserve order and time. A projector may also
derive a current locked/live amount and outcome state, but it must never mutate `S_i`.

Let `B^V_u(t)` be participant `u`'s available VITA capacity and `C^V_u(t)` its current carrying
capacity. A working bounded recharge family is:

`B^V_u(t + dt) = min(C^V_u(t), B^V_u(t) + rho^V_u(t) * dt)`

This expresses smooth renewable capacity without approving a recharge rate, time unit, cap, or
linear implementation. A saturating exponential or exact fixed-point equivalent remains available
for calibration.

Each later admitted evidence item `j` supplies signed alignment `a_ij`. A working aggregate is:

`x_i = sum_j I_independent(i,j) * w_type(j) * a_ij`

`I_independent` is zero for the stake owner, duplicate source, an unregistered oracle presented as
an oracle, or another source that fails the frozen admission policy. `w_type` distinguishes ordinary
independent evidence from registered-oracle evidence. Owner direction requires:

`w_oracle > w_ordinary > 0`

for otherwise equivalent admitted evidence, with the same absolute oracle weight for aligned and
contradictory evidence unless a later owner decision deliberately makes the system asymmetric.
The values remain experimental.

The current working bounded transformation is:

`F(x_i) = x_i / (K + abs(x_i))`

with `K > 0`. Positive `F` can increase promotion and future VITA flow. Negative `F` can increase
slashing and negative reputation. `K`, thresholds, and mappings from `F` to either consequence are
not canonical.

A working future-capacity expression is:

`A^V_u(t) = G^V(R^V_u(t))`

where `A^V` is future VITA replenishment and `R^V` is derived identity-sensemaking reputation.
`G^V` must be bounded. It changes future issuance, not the weight of VITA already staked.

## Evidence and replay boundary

The canonical history must distinguish at least:

- the immutable VITA stake;
- ordinary independent supporting or contradicting attestations;
- registered-oracle supporting or contradicting attestations;
- oracle registration and revocation state at the attestation's admitted order;
- subject and evidence identifiers;
- actor/control identity used for self-confirmation exclusion;
- stake slash, unlock, and future-replenishment entitlement outputs derived at a cutoff; and
- the parameter, admission-policy, semantic-model, projector, and state-schema identifiers.

Oracle attestations are append-only inputs. Projector/state-root oracles still perform cutoff-bound
replay and must converge on byte-identical canonical state. Evidence-producing oracles and
state-root attesters are separate roles even when one operator is eligible for both.

## Development replenishment bridge

The current Sepolia generation cannot implement smooth capacity or projector-issued dividends.
A temporary hosted development bridge may replenish an authenticated custodial wallet toward one
on-chain minimum vouch amount per UTC day, using a durable operation journal and the existing
development funder. It must:

- be labeled Sepolia development policy, not protocol economics;
- derive the temporary carrying target from the live contract's minimum vouch amount;
- transfer only the shortfall and never exceed that target through the hosted calculation;
- allow no more than one completed replenishment operation per participant per UTC day;
- preserve an exact intent, transaction hash, receipt, amount, recipient, and policy identifier;
- fail closed after an uncertain broadcast and never blindly retry; and
- be replaceable by projector-derived Base replenishment without changing historical VITA stakes.

This bridge fixes the unusable five-versus-ten development journey. It does not implement VITA
dividends, slashing, or Base production issuance.

## Calibration still open

- participant admission and independence/Sybil policy;
- wallet recharge function, cadence, half-life if any, cap, floor, and eligibility;
- stake live/locked lifetime, unlock rules, and whether reaffirmation is useful for VITA;
- ordinary evidence weight and registered-oracle extra weight;
- semantic alignment/extractor, noise floor, and duplicate-source rule;
- confirmation scale `K` and promotion/slash thresholds;
- partial versus full slash curve and where slashed units go;
- reputation aggregation and bounded `G^V` future-replenishment curve;
- issuance, concentration, and collusion ceilings;
- registered-oracle admission, revocation, bonding, and conflict policy;
- exact event schemas, ordering, fixed-point arithmetic, rounding, cutoff, and finality; and
- gas sponsorship, agent execution, and production claimant authorization.

## Required simulator comparisons

The VITA simulator must expose every input and cover:

- a five-VITA participant regenerating enough capacity for a ten-VITA minimum stake;
- recharge stopping at the configured cap;
- an immutable stake receiving later independent aligned evidence and a future-issuance dividend;
- ordinary versus registered-oracle aligned evidence, with the oracle producing more promotion;
- ordinary versus registered-oracle contradiction, with the oracle producing more slashing;
- self-support, duplicate support, and unregistered-oracle evidence contributing zero;
- one unit of VITA retaining one direct stake unit regardless of participant reputation;
- long-run future issuance potentially exceeding later validating stake while remaining bounded by
  the replenishment policy; and
- replay proving that an oracle appends evidence but does not directly resolve or rewrite history.

Every result remains `EXPERIMENTAL_NOT_CANONICAL` until an owner decision selects a complete
parameter manifest and a later release process supplies contract, projector, security, rehearsal,
and live evidence.

## Migration from V1

V1 correctly identified VITA as stake-backed identity work and anticipated both rewards and
slashing. V2 changes the active interpretation in four important ways:

- one-time acquisition becomes bounded renewable capacity;
- fixed rewards or transfers become future VITA issuance/replenishment dividends;
- oracle-panel claim resolution becomes append-only oracle evidence plus deterministic replay; and
- reputation multipliers on an existing stake are removed, while registered-oracle evidence gains
  a separate, symmetric, configurable evidence weight.

The V1 artifact remains unchanged below its supersession banner so design ancestry and deployed
development-contract behavior remain auditable.