EXP-06 · Evidence·Hemp'in Learn ▸ Library ▸ Ledger
Evidence Ledger
Every big hemp claim, graded in public. Each gets the version you hear, the version that was measured, a grade, its sources and the date someone last checked. Where the evidence is missing, the ledger names the study that would settle it and puts a bounty on it. Every release is a citable dataset.
01A grade exposes a gap
Each claim gets the version you hear, the version that was measured, a grade and its sources. A Weak row names the study that would settle it.
00One claim, two sentences
The grade belongs to the claim as people say it
“A hectare of hemp absorbs 22 tonnes of CO₂.”
Published estimates put it at 7 to 15 t per hectare in harvested biomass. And absorbed is not stored.
01The ledger
The public index
Nine claims seeded from a first evidence table, the honest starting set. Filter by area or grade, or select a band of the bar; open a row for what was measured.
▤hempin.org/learn/ledger
Published estimates put CO₂ absorbed in harvested biomass at roughly 7 to 15 t per hectare, depending on yield and practice. The 22 t figure is mostly advocacy. Absorbed is not stored.
02A claim page
Strong still comes with conditions
The unit of the ledger. The strongest row is the most useful one to design, because the page has to show both the grade and its conditions without hedging the headline.
◆Building materials · HC-01 · version 3
“Hempcrete walls are carbon negative.”
Hempcrete walls can store up to 16 kg CO₂ per m² net of producing them. Whether a given wall does depends on its mix.
- The binder is lime-led and the mix is light. A cement-heavy mix at 425 kg/m³ still emitted 8.16 kg CO₂/m² in the same study.
- The wall stands for decades. Storage counts only while the material lasts. The EU’s carbon-removal framework asks for at least 35 years.
- It is infill. Hempcrete is not structural; the frame’s carbon is counted separately.
| Arehart et al. 2020, quantifying the carbon uptake of hempcrete | peer-reviewed LCA via Carbon Leadership Forum | opened Oct 2026 |
| EU framework for carbon removals (CRCF), Delegated Act 2026/285 | legal summary · 35-year product storage | opened Oct 2026 |
| UNEP building-sector share of energy-related CO₂ | context, not evidence for the claim | summary only |
One harmonised life-cycle study per common mix, measured the same way, would replace “with conditions” with a grade per mix. That study is Bounty B3.
⌗How the grade was reached
A grade is the lowest of the five tests that matters for the claim as people say it. Try the rubric in the next section.
| Editor | Hempin research desk |
| Reviewer | Named academic partner to recruit |
| Conflicts | None declared. Hempin sells no hempcrete. |
Strong, with conditions. The EU 35-year rule added as a condition.
Hempin Evidence Ledger (2027). HC-01 Hempcrete stores carbon, v3. Release 2027.1. doi:10.5281/zenodo.[assigned on release]
Every submission and its outcome is logged on the claim, rejected ones included. Watching a claim earns no XP: nobody is rewarded for pushing a grade.
03How a grade is earned
Five tests, four grades, the lowest one wins
A myth often has a true measured version, and showing both is what makes the correction land. Load a row, or set the tests yourself and watch which one binds.
⚖Grade a claim · the five tests
◆The grade
Strong, with conditions
Set by Design and Agreement and Directness and Independence and Recency: the grade is the lowest of the tests that matter for the claim as people say it.
Two or more independent, peer-reviewed measurements agree, taken in conditions close to real use. Conditions are allowed and shown.
We publish, we do not define
Hempin keeps the record and the method. Standards stay with ASTM, ISO and the EU. A grade says what the cited evidence supports today, never what is “certified”.
Two keys to change a grade
An editor drafts. A named outside reviewer signs. No grade is published or moved with one signature, and both names sit on the page.
Claims, never companies
The ledger grades statements. It never names a company as the source of a myth.
No paid influence
Customers, sponsors and bounty co-funders cannot buy, delay or veto a grade. Every funder is listed. The fee buys the document, never the verdict.
Anyone can bring a source
Submissions are public, logged with their outcome and a reason, rejected ones included. A company that disagrees answers with evidence.
Checked on a clock
Strong rows are re-checked every 12 months, the rest every 6. Verra’s hemp reassessment (Q3 2027) triggers a re-check of every carbon row. An overdue row shows its grade greyed.
04BUD cites it
Answer first, grade second
The ledger becomes BUD’s source of record for every claim it covers: BUD states the grade and the check date, never repeats a myth as fact, and says plainly when the ledger has no grade yet.
✶BUD in Learn
Often, but it depends on the mix. A light, lime-based hempcrete wall can store up to 16 kg CO₂ per m² after its own production. A cement-heavy mix in the same study was still a net emitter.
Strong, with conditions · HC-01 · checked Oct 2026§BUD’s rules for ledger claims
- Answer first, grade second. The chip carries the grade so the sentence can stay plain.
- Lead with what was measured. When a question repeats a myth, BUD opens with the measured fact and names the myth once, at most.
- No grade, no verdict. For an ungraded claim BUD says the ledger has not reviewed it and labels what it found as unreviewed.
- Health is out of scope until a medical reviewer signs those rows. BUD says so and points to a professional.
- Every chip links to the claim page, so the answer and its sources are one tap apart.
BUD is grounded on 129 documents today. The ledger becomes its source of record for every claim it covers.
05Myth versus measured
The myth is never the last thing a viewer hears
Three of the ten episodes, storyboarded. Vertical, 45 to 60 seconds, one claim each, sources in the description, the claim page on the end card. The playhead runs while the board is on screen.
▶Ep 1 · 22 tonnes? · 55 s · vertical 9:16
It’s true: a hemp field pulls a lot of carbon out of the air while it grows.
06Evidence bounties
Pay for the answer, not the answer you want
A bounty is a public prize for a study the ledger says is missing. It is chosen by a rule, not by taste. Switch the rule’s tests to see what it keeps out.
⊞The selection rule
| Claim | Grade | Gap | Dated decision | ≤ 3 years | Why | Bounty |
|---|---|---|---|---|---|---|
| HC-01 Hempcrete walls are carbon negative. | Strong | ✓ | ✓ | ✓ | With conditions; EU life-cycle carbon disclosure from 2028; one LCA per mix fits in 3 years. | B3 |
| BU-02 Carbon stored in hemp buildings stays stored. | Weak | ✓ | ✓ | ✓ | Weak; Verra reassesses in Q3 2027; core samples from older buildings answer early. | B2 |
| CA-01 A hectare of hemp absorbs 22 t of CO₂. | Myth | — | — | ✓ | Graded Myth: the gap is in the telling, not in the evidence. | — |
| CA-02 Hemp farmers can sell carbon credits. | Myth | — | ✓ | — | A methodology decision, not a study, settles it. | — |
| TX-01 Hemp fabric uses far less water than cotton. | Medium | — | — | ✓ | Medium: supported, if old. Not the first gap to fund. | — |
| MA-01 Hemp is replacing plastic in cars. | Medium | — | — | ✓ | Medium; a market question more than a science one. | — |
| FA-01 Hemp boosts the yield of the next crop. | Weak | ✓ | ✓ | ✓ | Weak; CAP eco-schemes reward measured rotation; three seasons fit in 3 years. | B1 |
| SO-01 Hemp cleans contaminated soil. | Weak | ✓ | — | — | The evidence already answers it; no decision waits on more. | — |
| FR-01 Hemp batteries outperform lithium. | Myth | — | — | — | No study would change a decision now. | — |
3 of 9 rows qualify: HC-01, BU-02, FA-01.
◎hempin.org/learn/ledger/bounties/b1
Does a hemp break crop raise the yield of the cereal that follows, across sites and seasons?
CAP eco-schemes reward measured rotation and low-input benefits. Today one regional study carries the whole claim. Farm unions and policy briefs need a multi-site answer.
- At least 3 sites in at least 2 EU climate zones
- At least 3 seasons, randomised plots against a standard break crop
- Measure the following wheat yield, soil nitrogen and weed pressure
- Preregistered analysis; farm-level data deposited with a DOI
Weak FA-01 FA-01 from Weak to a grade the evidence earns, either way.
⋯Milestones, paid in stages
A null result is paid the same. The prize rewards a sound study that answers the question, whichever way it lands.
Preregistered first. Hypotheses, sites and analysis plan are filed publicly before data is collected.
Open data with a DOI. Open-access paper or preprint; data under CC BY 4.0 so anyone can reuse it.
An independent panel judges. Funders and Hempin sit outside it. Hempin gets no veto and no exclusivity.
Paid in stages. A share on an accepted preregistration, a share on open interim data, the rest on the final study.
Every funder is listed. Co-funders choose the question with the panel, never the result.
Amounts are set with co-funders. No prize figure appears until a co-funder has committed it.
Pledges wait for the right rail. A crowdfunded pool opens only once an invoice or SEPA payment route is live.
07Open datasets
Every release is citable
Each quarter the ledger is frozen into a release and deposited in a Hempin community on Zenodo, which mints a DOI. One concept DOI points to the newest release; each release keeps its own DOI forever, so a paper can cite the exact grades it read.
⛁Release 2027.1 · illustrative
Graded claims about industrial hemp with their sources, the five-test assessment, reviewers and full grade history. Sources are listed by metadata and link only; no text is copied from them.
- csvclaims.csvone row per claim version30 rows
- csvsources.csvone row per source per claim~90 rows
- csvassessments.csvthe five tests, per version30 rows
- csvgrade_history.csvevery grade change with both signersappend-only
- csvsubmissions.csvpublic source suggestions and outcomesappend-only
- jsondatapackage.jsonFrictionless schemaschema
- mdREADME.mdmethod, rubric, licence, how to citedoc
| Concept DOI | 10.5281/zenodo.[on first release] |
| This release | 10.5281/zenodo.[on release] |
| Licence | CC BY 4.0 for everything Hempin writes. Source metadata keeps its own licence, recorded per row. |
| Cadence | Quarterly, on the day Hempin publishes its claimed-pages count. |
⌗What a row holds
| File | Field | Meaning |
|---|---|---|
| claims | claim_id, version | Stable id (HC-01) and a version that increments on any change to text, grade or sources. |
| claims | street_claim, measured_claim | The claim as heard and the claim as measured, with units. EN first; FR and ES from the Learn translation pipeline. |
| claims | grade, conditions | One of four grades, plus the “true when” conditions as a list. |
| claims | checked_on, next_check, overdue | Dates, and a flag the moment next_check passes. |
| claims | gap, bounty_id | The study that would change the grade, and its bounty if one exists. |
| sources | doi_or_url, type, year, region | Type is one of: field measurement, LCA, review, model, industry report, advocacy, Hempin scan. |
| sources | opened, funding_disclosed | opened or summary_only. A summary-only source can never carry a grade alone. |
| sources | science_work_id | Link to the Library’s publication registry, so the ledger and the Library share one list of works. |
| grade_history | from, to, editor, reviewer, reason, at | Append-only. Corrections to Hempin’s own mistakes are logged here too, in public. |
08Research consortia
The data partner a consortium is missing
EU research consortia must publish their data openly and prove their results reach people. Most have a lab, a university and an industry partner, and nobody whose job is the data.
◇For a coordinator, one page
Evidence synthesis and open data for bio-based materials research.
We turn a project’s results into a graded, sourced public record that regulators, buyers and press can read, and into citable open datasets that meet the programme’s open-science rules.
- Data management plan and FAIR deposits: schema, metadata, DOIs, versioned releases.
- Evidence synthesis: the project’s claims graded against the literature with the ledger’s published method.
- Dissemination in EN, FR and ES: claim pages in Learn, BUD answers, short videos.
- Bounty protocols: a public, preregistered study design the consortium can run and the ledger will grade.
⟶Call pipeline
- nowCOST Action CA25133 (Legal Design), WG4A free working-group seat; the way into COST networks.fit: network
- 2 Feb 2027Horizon Europe CL4, textile circularityLedger rows on textile water and fibre claims; DPP context.fit: mediumin 78 days
- 11 May 2027Horizon Europe CL6, governanceAn evidence-to-policy record for CAP eco-schemes; pairs with the rotation bounty.fit: mediumin 176 days
- 22 Sep 2027CL6-2027-01-CIRCBIO-09, circular bio-basedOpens 20 Apr 2027. Traceability and data of bio-based resources.fit: highin 310 days
- 2027 work programmeCBE JU 2027 callPartner with research projects already studying hemp fibre.fit: highin 410 days
09Build spec
What it would take
One page per claim, a grade, sources, a check date, 30 claims, BUD citing them. The ledger touches public data, so it ships with row-level security and a public-safe read path.
↻The life of a claim
Two signatures
next_check passes, a trigger date arrives, or a submission is accepted
⌗The records
| Record | Holds | Who writes | Who reads |
|---|---|---|---|
| Claim | Stable id, area, slug, current version | Editor | Public |
| Claim version | Street and measured text, grade, conditions, gap, dates | Editor drafts; published only with a review | Public once published |
| Source link | Claim version ↔ Library work, opened or summary-only, type | Editor | Public |
| Assessment | The five tests, with a sentence each | Editor; reviewer comments | Public |
| Review | Reviewer, decision, reason, conflicts declared | Outside reviewer | Public |
| Submission | Suggested source, submitter, outcome, reason | Any signed-in user | Public after triage |
| Bounty | Question, protocol, status, milestones, funders listed | Admin | Public |
| Release | Frozen snapshot, DOI, date, files | Admin | Public |
◎How we would know it works
30 claims published and signed. BUD cites the ledger on every covered question. A first release with a DOI. Three bounties drafted, each with at least one named co-funder.
Releases cited in papers or consultations. Press links to claim pages. Accepted outside submissions. A grade that moved because someone outside Hempin brought evidence.
Over 20% of rows overdue. No outside reviewer signed after six months. A funder asks to see a grade before it is published.
!Risks, and the answer in the design
| Videos spread the myth they debunk | Measured fact first and last; the myth said once; titles state the measured fact. |
| An industry member feels attacked | Grades are on claims, never companies. Right of reply through evidence, logged in public. |
| Hempin gets a grade wrong | Corrections are a normal event in the grade history, published with the reason. A visible correction earns more trust than a silent fix. |
| Funders capture the agenda | The bounty selection rule is public; funders choose questions with the panel, never results; all funders listed. |
| The ledger goes stale | A check clock per row, an overdue state on the page, and a release cadence that forces a review pass each quarter. |
10Open decisions
Four calls the alpha leaves open
01 · Four grades, or five?
Lean: keep four (Strong, Medium, Weak, Myth) and carry nuance in “true when” conditions. A fifth “Unsettled” grade tends to become where hard calls go to hide.
02 · The first outside reviewer
Lean: ask one bast-fibre or building-materials lab to review the first ten rows. Without a named reviewer, nothing publishes.
03 · How bounties are funded
Lean: co-funders first (academic partners, foundations, grant lines that fund public goods); crowdfunding only after a second payment route exists.
04 · Licence
Lean: CC BY 4.0 over CC0, so every reuse carries a citation back to the ledger.