I have been arguing with a paper on my coffee table for the better part of a week.
It is a piece from Bright Meadow Group, passed along by way of Cernunnos Foundation, and it makes the case that intellectual property, as we have come to understand it, is a bottleneck on the species. I read it, set it down, picked it up, set it down again. I disagreed with it on first reading the way a man disagrees with weather — without much hope of changing anything, and with the suspicion that the weather has been at this longer than I have.
Property is the bedrock of the order I love. The republic the founders bequeathed us is, in no small part, a defense of what a man may call his own — his land, his labor, the fruits of his thought. To touch that is to touch the foundation. So I bristled. I do not mind admitting it.
But the founders had a habit, when they bristled, of sitting down to it. They wrote letters. They read what their adversaries actually said, not what their faction said their adversaries said. They examined the matter the way Grammina examined a pie crust — knuckle to the dough, eye on the seam, no opinion offered until she’d looked under the rim.
So I did the work. I went back to Jefferson.
He has a letter — to Isaac McPherson, in the summer of 1813 — in which he says, in his way, that an idea is a strange sort of property: that if I light my candle at yours, you are no darker for it. Grammina had a version of this. She told me once, when I was perhaps eleven and arguing with my cousin over a comic book, “You can light a man’s candle without going dim, Nate. Don’t make property of what doesn’t burn down.” I thought, then, it was about sharing. I see now it was about something larger.
It was about the difference between scarcity and enclosure.
A field is scarce. Two men cannot plow the same furrow. But an idea is not a field. It is closer to a song. And the law we built for fields, we have, by long habit and longer paperwork, draped over songs. The fit is poor. It has always been poor. Jefferson said as much when the ink was still wet on the country.
I owe it to the man whose republic I keep trying to deserve to read the case against IP not as an attack but as an examination. So here is the paper as they wrote it. I have made no edits. I will offer my verdict after.
Open Knowledge as Infrastructure
A Practical Framework for Advancing Machine Learning and the Humanities Timeline
Bright Meadow Group — Systems Analysis and Solutions Consulting
I. Problem Statement
Fragmented Knowledge is Now a Structural Bottleneck
Human knowledge is not scarce. It is fragmented.
- Locked behind paywalls
- Trapped in municipal archives
- Buried in university repositories
- Obscured by copyright regimes
- Duplicated across institutions
- Detached from provenance and context
This fragmentation made sense in a slower era. It does not make sense now.
Machine learning systems require breadth and contextual grounding. Human societies require historical continuity and traceability of ideas.
Instead, we operate inside silos that:
- Reinvent solved problems
- Lose minority or regional insight
- Reward ownership over synthesis
- Prevent adjacency-driven discovery
The cost is not academic inconvenience. It is civilizational drag.
We are attempting to solve global-scale problems with partitioned information structures.
II. Why Now
Three forces converge:
1. ML Systems Are Context-Hungry
Modern models can synthesize across domains — but only if those domains are accessible, structured, and legally usable.
Without comprehensive corpora:
- Evaluation becomes biased
- Historical nuance is flattened
- Cross-disciplinary synthesis collapses
- Alignment discussions lack real context
Machine learning does not just need more data. It needs connected data.
2. Humanities Require Recoverable Lineage
The humanities timeline is fractured. We know conclusions. We often lose the argument paths.
Imagine:
- Every pamphlet,
- Every municipal record,
- Every research note,
- Every early draft,
- Every marginalia,
- All linked by canonical IDs.
The result is not nostalgia. It is traceable intellectual evolution.
3. Global Problems Are Networked
Energy, climate, governance, logistics, food systems — these are cross-domain problems.
Siloed knowledge cannot solve networked crises.
Synthesis is now the bottleneck.
III. Objections — Steelmanned
Objection 1: “Creators must be paid.”
Correct.
Access control is not the only compensation model.
Alternative models:
- Public knowledge dividends
- Usage-based pooled payouts
- Direct public grants
- Institutional subscription support to creators (not readers)
- Prize-based reward structures
- Voluntary patronage with open release requirements
Compensation and enclosure are separable variables.
Objection 2: “What about misuse?”
High-risk operational domains (e.g., certain biotechnologies) require tiered governance.
However:
- 95% of human knowledge is not dangerous.
- Suppression does not eliminate malicious actors.
- Transparent systems are easier to monitor than fragmented ones.
The proposed system:
- Does not execute decisions.
- Does not control infrastructure.
- Does not automate action.
It surfaces, links, audits, and contextualizes.
Objection 3: “Intellectual property incentivizes innovation.”
Historically true in specific industrial contexts.
However:
- IP systems were designed for physical scarcity.
- Knowledge is non-rivalrous.
- Enclosure slows distributed innovation.
Innovation today is increasingly combinatorial. The value lies in synthesis, not exclusivity.
Objection 4: “Privacy and consent.”
Non-negotiable.
Open knowledge does not include:
- Private medical data
- Personally identifying data without consent
- Community-restricted cultural materials
Privacy protections are architectural requirements, not afterthoughts.
IV. The Open-Knowledge Stack
A layered architecture makes this feasible.
Layer 1: The Corpus
- Public-domain-first ingestion
- Voluntary rights-holder participation
- Canonical identifiers for:
- Works
- People
- Institutions
- Places
- Concepts
- Version-controlled documents
- Deduplication
This layer is the substrate.
Layer 2: The Timeline Graph
Every entry becomes a node:
- Event
- Claim
- Publication
- Person
- Institution
Edges define:
- Influence
- Citation
- Dispute
- Collaboration
- Context
Competing interpretations are preserved, not resolved. Provenance is explicit.
This is the humanities engine.
Layer 3: ML Integration Layer
- Transparent training datasets
- Evaluation benchmarks with contextual grounding
- Auditable filtering policies
- Model comparison against historical knowledge graphs
This prevents context collapse and hallucination loops.
Layer 4: Governance
Principles:
- No centralized ideological control
- Independent oversight
- Public audit logs
- Abuse monitoring
- Tiered access only where objectively necessary
The system must never have operational authority.
It is a map, not a lever.
V. A 24-Month Pilot (Doable, Realistic)
Scope small. Prove value.
Phase 1 (Months 1–6): Public-Domain Core
- Federal documents
- State archives
- Public-domain books
- Local newspapers (pre-1928 US)
- Open-access academic journals
Build canonical IDs and graph structure.
Phase 2 (Months 6–12): Municipal + Regional Integration
Select a specific region (e.g., a mid-sized American city with historical archives).
Integrate:
- Municipal meeting minutes
- Zoning records
- Infrastructure development plans
- University thesis repositories
- Local newspapers (open years)
Construct a complete 150-year regional knowledge graph.
Phase 3 (Months 12–18): ML Interface Layer
- Build query tools
- Build adjacency visualization
- Train limited models exclusively on this open stack
- Compare outputs against proprietary-trained models
Measure:
- Reduction in hallucinations
- Cross-domain synthesis quality
- Historical reasoning performance
Phase 4 (Months 18–24): Public Release + Academic Partnerships
- Open API
- Research competitions
- Humanities timeline visualizations
- Public dashboards
Demonstrate that aggregation + organization + open parsing produces measurable gains.
VI. What Emerges
When adjacency increases:
- Lost ideas reappear.
- Parallel inventions become visible.
- Suppressed contributions surface.
- Pattern recognition accelerates.
This is not speculative optimism.
Network theory predicts it.
VII. Closing Position
We do not give such a system control.
We do not automate governance.
We do not centralize authority.
We build a shared substrate that:
- Surfaces
- Links
- Preserves
- Audits
- Contextualizes
Knowledge enclosure made sense when information moved slowly.
It does not make sense in a world where synthesis is survival.
I have read it twice now. I have argued with it twice. The argument has not gone well for me.
What strikes me, on the second reading, is that I am not being asked to surrender property in the way I feared. I am being asked to see that we have called two different things by the same name. There is property that diminishes when shared — the field, the loaf, the roof over my head. And there is property that compounds when shared — the proof, the poem, the technique for setting a bone. We have, for two centuries, treated the second as if it were the first. The cost of that mistake is what these people are trying to count.
I am a Jeffersonian. So, as it turns out, was the case I was trying to refute. That is humbling, but it is also instructive: a man who reads the founders carefully should not be surprised to find them on the other side of a fight he assumed was settled.
The objections they steelman are mine. Creators must be paid. Yes. The paper does not deny this; it separates payment from enclosure, which is a separation I had not thought to make. What about misuse? Yes — and they concede tiered governance where it is honestly needed. Privacy. Yes. Non-negotiable, and they say so plainly. Incentive to innovate? Here I lingered the longest. But the closer I looked, the more I noticed that the loudest defenders of IP today are not the inventors. They are the holding companies. Jefferson would have had a word for that. Grammina would have had a shorter one.
So I find myself, at the end of a week of arguing, conceding the field. Not all of it — I will hold my line on privacy, on consent, on the dignity of the working creator — but the central enclosure, the habit of fencing the candle’s light, I can no longer defend on republican grounds. The republic is sustained by citizens who think for themselves and pay the cost of doing so. I owe the reader the cost.
We were taught that property is the foundation of liberty. It is. But the foundation has to be set in the right soil. Songs do not grow in fields, and fields do not grow in songs. Build each its own house.
Grammina, when I told her once that I had changed my mind about something I had argued for years, did not congratulate me. She said, “That’s not changing your mind, Nate. That’s catching up to it.”
I am catching up.
— N.L.
