Can continuity exist independently of identity?

Background

Identity has traditionally been treated as the prerequisite for continuity. To verify that a subject has maintained continuous existence, we first ask who the subject is — a wallet address, a DID, a stored template. Continuity, in this view, is a property that identity enables.

The Continuity Lab explores the inverse hypothesis:

If this hypothesis holds, it would invert the current relationship between identity and continuity. Today, identity is the gatekeeper and continuity is an optional extension — checked only when something goes wrong. Under the inverse model, continuity is the substrate: every digital subject continuously proves its existence, and identity labels are applied post-hoc to trajectories that meet certain criteria.

Why It Matters

If continuity can be verified without relying on persistent identifiers, three consequences follow:

1. Humans and AI agents share a common verification substrate. A human proving continuous presence and an AI agent proving continuous operation use the same primitive — continuity proofs. The distinction between "human identity" and "agent identity" becomes a policy layer on top of a shared cryptographic foundation, not two separate protocols.

2. Identity becomes revocable without breaking continuity. Today, if you lose your private key or your DID is compromised, your continuity is also lost — because your identity was the anchor. Under a continuity-first model, you could rotate your identity credentials while maintaining the same continuity trajectory. The trajectory is the permanent record; the identity labels are replaceable.

3. Digital twins and hybrid entities become tractable. A human delegates a task to their AI agent. The agent operates for two weeks, then returns. Under the identity model, we ask: "Is this the same agent?" Under the continuity model, we ask: "Has the trajectory been unbroken, and can we trace the delegation chain?" The second question is both more precise and more general.

Current Evidence

Conceptual only. No empirical validation yet.

The PES Benchmark (BM-001) demonstrates that human presence can be distinguished from synthetic motion with high confidence (Cohen's d = 2.1, AUC = 0.94). However, this establishes presence detection, not continuity verification. The gap between "this is a human right now" and "this human has continuously existed across the interval" is the central open question.

RN-001 (The Continuity Problem) frames the theoretical case: identity succeeds and continuity fails in four specific attack scenarios. But a theoretical case is not an experimental result. The next step is to move from "here is why this matters" to "here is how we test it."

Next Milestone

Define a formal Continuity Model — a mathematical specification of what it means for a digital trajectory to be continuous. This model must:

1. Define the minimum sampling frequency required to establish continuity across a time interval.

2. Specify the conditions under which a gap in the trajectory constitutes a continuity violation (versus a benign network interruption).

3. Provide a formal definition of "sovereign existence" — what properties must be preserved across the trajectory for continuity to hold.

4. Be falsifiable: there must exist an experimental setup that could demonstrate the model is incorrect or incomplete.