Why I Work in the Open
A note on method, validation, and where to find everything.
People sometimes ask why a methodology this developed a 500-scene annotated dataset, a rule-based detector, a pre-registered validation protocol is not sitting in a peer-reviewed journal. The honest answer is that the journal route was never the right fit for what this work is and what it needs, and choosing the open route was a decision, not a fallback.
Here is the reasoning, stated plainly.
The work is built to be checked by others, not by me
Objective Projection makes a falsifiable claim: that a defined matrix of physical, sensory variables raises the statistical probability of a convergent autonomic response across readers. It does not claim an identical response in everyone — the position was revised from determinism to probabilistic convergence precisely because individual variation is real and has to be built into how the claim is stated.
A claim like that means nothing if I measure it myself. Self-measured validation of one's own theory carries no evidential weight, and it shouldn't. So the validation protocol (OPCT v2.0) is published and pre-registered specifically so that independent researchers can run it — ECG, GSR, pupillometry, the whole instrument specified in advance, with no room for me to adjust the target after seeing the data. The map is drawn and the instrument is laid out on the table. It now needs other people's hands to take the readings.
This is the part the traditional pipeline handles poorly. A journal validates a finished result. This work is at the stage of a specified, open instrument waiting for replication. Those are different things, and open archiving serves the second one better than a closed review cycle does.
The pace is real
The systems this methodology studies large language models change on the order of weeks. A review cycle that takes a year is not a neutral delay for this subject; by the time it clears, the object of study has moved. Putting the full corpus in the open, under a license that lets anyone inspect and audit it, keeps the work in step with the thing it describes.
Everything is already public, and that is the point
There is no hidden version. Every piece of this project the dataset, the detector, the protocols, the conceptual papers, the version history — is openly archived and listed in one place: the corpus index at leventbulut.com/corpus/. The primary dataset lives on Hugging Face (leventbulut/objective-projection) with a permanent archival mirror on Zenodo, and the validation protocol is registered on OSF (osf.io/us8bw).
Anyone is free to read it, teach from it, build on it, critique it, or attempt to falsify it. The only thing the license withholds is commercial reuse. Beyond that, the door is open — including to researchers inside the journals. If a finding here is useful to your work, use it; that is what it is here for.
What I am not claiming
I would rather state the limits than have them found for me.
The evaluation-layer claim that an LLM acting as a judge re-applies an emotional label internally and penalizes text that correctly withheld it is at present a hypothesized mechanism, not a measured result. The narrative-entropy operationalization has a known dimensional-consistency problem that the pilot report states openly rather than papering over. And the convergence claim itself is, as above, exactly the thing that independent measurement is meant to confirm or break.
None of this is settled science presented as settled. It is an open, auditable body of work with its uncertainties marked on the surface which is, I think, the most honest place for it to be.
The complete corpus, with every document and its archival identifier, is at leventbulut.com/corpus/. The dataset is at huggingface.co/datasets/leventbulut/objective-projection. The validation protocol is registered at osf.io/us8bw.
