DvF is an open, collaborative research project investigating whether physical law determines unique outcomes or instead determines structured constraints within which outcomes occur. It brings philosophical analysis, formal modeling, literature review, experimental design, software, and reproducibility records into one assumption-indexed workspace.
Project status: DvF is an open research project in its formative stage. The repository currently contains a prospectus, methodological framework, proposed experiments, and collaboration infrastructure. It does not yet contain completed physical experiments or evidence establishing metaphysical indeterminism.
Read the preserved full research prospectus for the project's motivating argument, proposed experimental architecture, qualifications, and scope.
The core research question is: Does physical law select one unique outcome, or can it specify a structured range of admissible outcomes without selecting the realized outcome in advance?
The project calls the latter possibility constrained openness: lawful constraints can govern possible outcomes, probabilities, and correlations without themselves implying that every individual result was uniquely selected beforehand. This is a hypothesis space, not a settled conclusion. DvF does not presume that metaphysical indeterminism has been proven; it tests model classes and makes every inference conditional on its assumptions.
Contributions are welcome in philosophy, physics, mathematics, statistics, experimental design, software, replication, and transparent AI-assisted analysis.
- Contribution guide and governance
- Instructions for AI agents
- Documentation site source
- Research records
- Experimental arms
- Roadmap
Python 3.11 or newer is required.
python -m venv .venv
source .venv/bin/activate
python -m pip install -e '.[dev]'
make allRun only repository validation with python -m dvf validate --strict or make validate. Build the documentation locally with make docs.
Substantial AI-assisted contributions are disclosed and recorded under agent-runs/, including inputs, outputs, verification, limitations, and handoff notes. These records document decisions and evidence, not private chain-of-thought. AI output receives human and technical review and is never a substitute for evidence.
Licensed under the existing Apache License 2.0.