From Third- to First-Person Science

The third-person face

Most of what we do at BCOM is third-person science. We build computational models of agents—their world models, their valences, their plans—and we use those models to predict what brains and behaviors will do. The program has a name: computational neurophenomenology. It feeds directly into computational neuropsychiatry, neurotwins, and the clinical agenda.

The arc is the familiar one of science. Build a model. Predict the data. Compare. Iterate. The data we want to predict includes a subject's verbal report about their experience: their pain, their presence, their mood, the quality of a hallucination. If our model of the agent cannot predict the report, the model is incomplete as a model of mind.

The philosophical undercurrent

Underneath the third-person work, we hold a stronger claim. In the Kolmogorov Theory framework, a sufficient computational model of an agent does not merely predict what the agent will say about its experience. It predicts the experience itself.

The conceptual scaffolding for this claim is the Unicum.

Under this view, computation is not a process layered on top of a dead substrate that somehow lights up into qualia. Where the experiential field is organized into agent-shaped structure—a modeling engine, an objective function, a planning engine—structured first-person experience is what that organization is, from the inside. A good computational model and the lived experience are two faces of one ground.

This is metaphysics, and we hold it as a working stance rather than a proven thesis. But it changes what we ask of our models. We do not only want them to predict reports. We want them to predict experience.

First-Person Science

That ambition forces a methodological move. If a model claims to predict your experience—not your behavior, not your verbal report, but what it is actually like for you—then you are the natural place to test it. Perturb the agent's computational state (with neuromodulation, pharmacology, contemplative practice, task demands), let the predicted experience play out, and check it from the inside.

We call this First-Person Science. The researcher becomes both the experimental subject and the experimental judge. Validation happens in the experiential substance of the scientist. Then the result is written up in the usual third-person way—methods, predictions, observations, comparisons—and submitted to peers, who run the same protocol on themselves and converge, refine, or diverge.

The intersubjective machinery of science is preserved. What changes is the locus of the measurement: from a verbal report about an experience, to the experience itself, witnessed by the only instrument that has direct access to it.

The three together

The BCOM Philosophy → Brains pillar runs the three at once. We do third-person computational neurophenomenology and neuropsychiatry. We hold the Unicum as conceptual scaffolding, so that what our models predict can in principle be experience itself, not only its 3P shadow in a report. And we develop First-Person Science as the methodology that turns this from philosophy into practice.

It begins as third-person science. It ends as a new kind of science.