THE RESEARCH ENGINE

Everything else on this site describes PSYCHO-VULGARISM as a theoretical and historical framework — drawn from psychoanalysis, neuroscience, and the convergent testimony of every major initiatic tradition. This page describes something different: the ongoing, automated effort to confirm and refine that framework against real, unscripted human language, gathered continuously from the actual places where the hive mind speaks.

The goal is not to prove PV exists in the abstract. It is to build a working, empirically-grounded map of exactly how it sounds — precise enough that the pattern becomes recognizable in any conversation, on any platform, regardless of which political tribe is speaking.

What Is Being Collected

The research engine continuously gathers public conversations from multiple online platforms — political discussion boards, long-form debate forums, and public comment sections beneath contested video content. Nothing collected is curated, staged, or solicited. The entire premise of the project depends on capturing genuine, unscripted responses to genuine triggers — the way people actually argue, deflect, and defend when their identity feels threatened by an idea, not a simulation of how they might.

As of this stage of the project, the archive holds more than 8,751 collected transcripts, spanning a deliberately broad range of contested topics — political, social, and ideological — specifically because Psycho-Vulgarism is not confined to one side of any argument. The hive mind does not have a party. It has a mechanism. Capturing that mechanism requires observing it wherever identity-protective language actually appears, across the full spectrum of what people are currently arguing about.

Growth Over Time

The three counts above grow at deliberately different rates. Collection is largely automated and runs continuously. Full analysis against the Logic file takes real processing time per transcript (see Hardware and Scale, below), so the processed count climbs far more slowly. And the Logic v2 count starts from zero — every transcript has to be re-analysed against the new rubric. Those gaps are expected, not a sign anything is stalled.

Scraping began August 9, 2026. Processing began August 10, 2026. The chart plots measured daily snapshots beginning August 30, 2026; earlier counts were not recorded.

The Logic File — How the Analysis Works

Every collected transcript is passed through a structured analytical framework internally referred to as the Logic file. This is not a keyword search or a sentiment score. It is a detailed behavioral rubric of engine scoring markers — the criteria the engine scores each transcript against, seven in Logic v1 and five in Logic v2 — built from the Psycho-Vulgarism framework described elsewhere on this site, and separate from the four observation markers on the Psycho-Vulgarism page, which are written for reading a live exchange. They include the reflexive weighting of who said something over what was said, the fluency of a defense that arrives faster than the challenge that provoked it, the disproportion between a stimulus and the emotional response it triggers, and the other observable signatures of a drive running in its archonic, defended direction rather than its biophilic, engaged one.

A consequence worth stating: because these are directions rather than states, "restoration" never means converting a person from one type to another. It means a specific drive changing direction — which can happen to one drive while the other six are unaffected. THIS FRAMEWORK'S OWN SYNTHESIS

An AI system trained against this rubric reads each transcript and scores it against every engine scoring marker — not asking whether a position is correct, but whether the manner of defending it shows the specific behavioral fingerprint the framework predicts. The output is structured, quantified data: which markers activated, how strongly, and under what kind of provocation.

The First Batch

Over 1,400 transcripts have been processed to completion. That batch was not intended as a result. It was a test of whether the pipeline holds end to end — collection, segmentation, analysis, storage — at volume and unattended.

It held, and it produced something the design phase could not: the first pattern data drawn from real analysed output rather than from the framework's own predictions. Some of what the rubric expected to find appeared exactly as specified. Some did not appear at all. Some patterns appeared that had not been anticipated.

That output is what Logic v2 was built from.

Refinement — How the Framework Improves Itself

The Logic file is not fixed. It is a living document, revised as the accumulating data reveals where the framework's current predictions hold and where they need sharpening. This is a genuinely empirical process, not a top-down theory being confirmed by cherry-picked examples — patterns that repeat reliably across thousands of independent conversations get reinforced into the framework; patterns that turn out to be noise, or that the framework mismeasures, get corrected.

Logic v1 was the framework's predictions written as a rubric. Logic v2 is the first version corrected against real output. The changes came from the 1,400-transcript batch: markers that fired reliably were sharpened, markers that fired on noise were tightened or cut, and patterns the first rubric had no category for were given one.

This is the difference between a theory of Psycho-Vulgarism and a working detector for it. Logic v2 is now being applied to the archive. A v3 should be expected, on the same basis.

Hardware and Scale

The analysis now runs on a Quadro RTX 8000 with 48GB of video memory, replacing the modest hardware the project started on. The model is Ministral 14B, running locally.

Local matters more than raw speed here. There is no per-token cost, no rate limit, and no third party holding the corpus. The machine can run continuously, and the analysis can be re-run against a revised rubric without paying for the same work twice — which is what makes iterative refinement of the Logic file practical rather than theoretical.

The constraint is no longer compute. It is time. Full analytical depth takes tens of minutes per transcript, against a collection pipeline that gathers on the order of a thousand a day. The gaps between the raw count and the two processed counts above are a function of that arithmetic, and they will widen before they close.

What Patterns This Will Reveal

The working hypothesis — already well-supported by the convergent testimony documented across the rest of this site, from Fromm's AUTOMATON CONFORMITY to Reich's CHARACTER ARMOR to the neurological architecture described in the NEUROSCIENCE section — is that Psycho-Vulgarism has a genuine linguistic signature. Not a set of political positions, but a detectable pattern in how positions get defended: the specific rhythm of reflexive, pre-loaded response that shows up regardless of which side of an argument a person occupies.

If the hypothesis holds against a large enough, sufficiently diverse body of real discourse, the payoff is significant: a demonstrated, data-backed answer to the question of whether the hive mind is a real, recognizable phenomenon in language itself — not a rhetorical accusation one side levels at another, but a mechanism that operates identically across the political spectrum, visible in the data regardless of which tribe is speaking.

Publication

The intent of this project has always been public verification, not private conclusion. Once the collection archive and the Logic file have matured to a scale and reliability worth standing behind, the underlying research corpus — the collected transcripts and their structured analysis — will be made available for independent review. Anyone will be able to examine the same data this framework is built on and judge for themselves whether the pattern holds.

This page will be updated as that milestone approaches. For now, the work continues — one conversation, one transcript, one refinement at a time.