Codexarium
A herbarium for texts: every specimen sourced, every disagreement kept.
Codexarium is an open, non-profit data commons and spatial knowledge platform for biblical and ancient south-west Asian text scholarship. It is the largest of the things I am building, and the one where the constraints matter most.
The organizing metaphor is a herbarium. In a herbarium, you do not paraphrase a plant; you pin the actual specimen to a sheet, label it with where and when it was collected, and keep it. Codexarium works the same way with texts. Contributors deposit originals, preserved verbatim, with provenance labels attached. Anything derived from those originals — a reading, a translation, an argument — is recorded separately and points back to the pinned specimen. The observation and the source it rests on never get collapsed into each other.
That separation is the whole point. Every assertion in the system carries three things: where it came from, what validation state it is in, and where it sits on a consensus tier. No unsourced, model-generated prose is ever published. If something cannot be traced, it does not go in.
Keeping disagreement rather than resolving it
The capability I care most about is comparison across frameworks. Take one passage that four scholars attribute four different ways. The usual move is to pick a winner and flatten the rest into a footnote. Codexarium keeps all four positions, side by side and queryable, so the disagreement is represented rather than erased. Scholarship is mostly argument, and a tool that hides the argument is lying about the field.
This is also the axis the commercial products cannot occupy. Logos and Accordance are closed by construction. Codexarium is open, computable, reproducible, attributed, and free — not as a marketing posture but as a structural choice, because a commons that you cannot read, query, or check is not a commons.
Where this goes, longer term, is a research institute for ancient text studies, run as a working model of human and machine cooperating on the same evidence under the same rules. The full project lives at codexarium.org.