PKnet · Academics
An instrument that behaves like one, and a seat inside its community.
PKnet Academics is the researcher-facing surface of the Peacekeeper Network. Custody-preserving access to the measurement infrastructure, a bounded annual compute allocation, and a published seat in the methodology of the field the instrument helps constitute.
- access
- instrument
- custody-preserving, glass-box
- allocation
- per team
- bounded, annual, equitable
- default
- non-commercial
- shareable, publishable, citable
- contribution
- signed
- to the shared research commons
What academics receive
Four things, in this order.
- 01
The instrument.
Access to the measurement infrastructure as researchers, not as consumers. A deterministic, glass-box measurement of interaction structure that behaves the same way in your lab as it does in ours.
- 02
A bounded annual allocation.
A per-team compute allocation, provisioned behind the instrument. Enough for a serious research programme. Bounded, so that access remains equitable across teams and jurisdictions.
- 03
A protocol registry, to cite and extend.
A public, append-only registry of released banks, calibration artefacts and signed protocol updates, versioned and citable. Reproducibility is the default posture, not an aspiration.
- 04
Methodological standing.
A published seat inside a working measurement community. The methodology that comes out of the network is co-produced. When PKnet's reading becomes the default reading, the academic contribution is on the record.
What academics carry
Custody, citation, non-commercial by default, contribution to the commons.
Custody preserved.
Custody of the record stays with the person whose interaction was measured, or with an independent civic custodian. Access to the instrument does not include access to identified data. This is architectural.
Non-commercial by default.
Findings are shareable, publishable, and open by default. Any commercial application of the same instrument moves to a separate seat and a separate contract, so that the academic commons is not lightly reassigned.
Cite and endorse.
Publications cite the doctrine and the specific protocol version used. Endorsement of the doctrine is separate from a paper's conclusions, and both are welcome to disagree, in public, with the network.
Contribute to the commons.
What one team learns to recognise becomes an input to the shared protocol registry. Others can replicate it, extend it, or contest it. Progress compounds because contributions are signed.
Shapes of academic work
Three ways a research team can engage the instrument.
Same instrument, three research shapes. Each shape holds the same custody invariants and yields reproducible, citable outputs.
Measure existing corpora.
The most direct shape. A research team applies the instrument to an interaction corpus (public, licensed, or institutionally custodied) and reports the reading. Because the instrument is deterministic and the bank is versioned, another team can replicate the reading exactly. The output is publishable and citable in its own right.
Measure a controlled interactive setting.
A research team designs a controlled, interactive setting inside PKnet's physics-based world model, runs it with a consented cohort, and measures the interactions inline. The instrument runs underneath; participants keep custody of their record; the paper reports population-level structure, not individual behaviour.
Compose the two above.
The two shapes composed, so a controlled setting is calibrated against readings on existing corpora and vice versa. Useful when a research programme needs to move between induced and observed interactions to isolate the structure the paper is claiming to measure.
The protocol registry
Everything you cite, published in the open, versioned, and reproducible.
Signed, versioned, replicable
The instrument is the same in your lab as in ours.
Every released bank, every calibration artefact, every protocol update is written to a public, append-only registry. Determinism is a discipline of the network, so a reading you publish today can be replicated by an adversary tomorrow, on the same version, and the answer will be the same.
Reproducibility is not aspirational. It is a property of the artefact you cite.
Who is in the network
Four kinds of institution.
- research-intensive
- universities
- Russell Group, U15, Ivy Plus, C9
- policy schools
- public affairs
- governance, technology policy, safety
- field labs
- independent
- safeguarding, cyber, public interest
- clinical + civic
- applied
- mental health, welfare, education
Enter the network
One page. Question, cohort, custody, publication plan.
Applications are read against the doctrine, by the Blankstate team that runs the programme. The same standard is applied everywhere. What matters is the question, the cohort, and the plan for custody and publication.
We reply within the week. If the fit is right, the first conversation is about the question, not the paperwork.
What the application says
- 01 The research question, in one paragraph.
- 02 The interaction cohort, how it is recruited, and the custodian holding the record.
- 03 The shape of engagement (A, B, or C above), and why.
- 04 The publication plan and the venue.