Public discourse no longer fails only when people lack information. It fails when claims, evidence, repetition, incentives, emotion, and institutional authority collapse into one indistinguishable blur.
Poliscope is building the evidence layer for contested public discourse. At its core is PEM, the Poliscope Epistemic Model: our proprietary methodology for tracking how claims emerge, travel, mutate, and harden into belief.
Poliscope does not pretend to be an oracle. It does not replace judgment. It does not turn contested reality into a single magic score. Instead, it gives analysts, journalists, researchers, institutions, and decision-makers a structured way to preserve the evidence trail.
Public Knowledge Has No Memory
Traditional fact-checking usually starts after a claim has already travelled. That work is valuable, but it often arrives after the public narrative has already moved on.
Claims do not move as isolated statements. They move through networks of sources, platforms, incentives, images, fragments, official denials, partisan interpretations, market reactions, search spikes, and social repetition.
Public discourse needs a memory layer: a way to preserve not only the current assessment, but the trail of how that assessment was reached.
What Poliscope Separates
A claim is not the same as a narrative. A source is not the same as an actor. A signal is not proof. Amplification is not corroboration. A prestigious outlet can repeat a weak claim. A weak source can publish a real receipt.
Claims
The specific assertions being made.
Receipts
The posts, articles, documents, images, videos, statements, datasets, or archives attached to those claims.
Sources
The accounts, outlets, institutions, or platforms that publish or amplify claims and receipts.
Actors
The people, organizations, governments, networks, or institutions named in or relevant to the claim.
Narratives
The larger storylines through which claims gain meaning, emotional force, and political utility.
Signals
Changes in the information environment, such as search interest, market movement, source volume, polling shifts, social pickup, or repetition patterns.
Assessments
Provisional judgments about what is supported, disputed, unresolved, or changing.
Reviews
Human or system updates that preserve why an assessment changed over time.
What Poliscope Refuses To Confuse
Amplification is not corroboration.
If ten outlets repeat the same unsupported origin, the claim has become more visible, not necessarily better evidenced.
Consensus is not corroboration.
Agreement can reflect independent convergence, but it can also reflect copied reporting, shared incentives, institutional caution, or one claim travelling through many mouths.
A signal is not proof.
A market move, search spike, or social surge can show that the environment around a claim has changed. It does not, by itself, explain why.
Narrative utility is not factual support.
A claim can be useful to a storyline whether it is true, false, exaggerated, or unresolved.
Source reputation is not truth.
The relevant question is not only who said it, but what evidence they provided and whether other evidence independently supports it.
Corrections are part of the record.
A serious system must preserve retractions, disputes, contradictory receipts, and changes in confidence.
When Agreement Outruns Evidence
Consensus does not prove that a claim is wrong. It also does not prove that a claim is right. Confidence should rise when independent methods, observations, and receipts converge—not merely when more institutions repeat the same proposition.
Poliscope calls it Consensus Creep when an unverified claim appears to gain institutional certainty faster than independent evidence accumulates. Qualifications may disappear, repeated language may harden, and the existence of widespread coverage may begin to substitute for the missing receipt.
Consensus Creep can emerge organically from deadline pressure, professional incentives, source dependence, or social conformity. It can also be consistent with narrative laundering, coordinated amplification, or other manipulation. The signal is therefore an instruction to trace the evidence chain—not proof of deliberate coordination and not a verdict on the claim itself.
A claim can ultimately prove true and still exhibit Consensus Creep if institutions adopt the conclusion faster than they independently establish its evidentiary basis. The inverse matters too: weak institutional uptake does not disprove a well-evidenced claim. Attention, professional incentives, geopolitical salience, and narrative utility shape which findings travel.
PEM calls the resulting comparative pattern Attention Asymmetry: materially comparable claims receive radically unequal scrutiny, repetition, or institutional urgency. The asymmetry may reveal differences in access, incentives, audience demand, political utility, or organized amplification. It does not prove that the prominent claim is false, or that the neglected claim is true; it identifies an allocation of attention that itself requires explanation.
For grave legal or moral classifications, PEM therefore counts independent applications of the relevant standard—not scholars, signatures, articles, or institutions as if agreement itself were a receipt. The question is whether separate analyses reach convergence through evidence, or whether one conclusion is acquiring authority through repetition.
PEM also separates epistemic proximity from social authority. Direct observation, primary evidence, and relevant field competence may carry more evidentiary value than prestige or popularity, while no witness, expert, institution, or title is treated as infallible.
Count independent evidence paths, not agreeing voices.
From Claim To Auditable Record
A Poliscope record begins with a claim, link, post, article, screenshot transcript, or analyst note. The system structures the object so it can be reviewed over time.
When The Claim Is Not Just A Claim
Some public arguments do not make a narrow factual claim. They recode ordinary conduct, institutions, identities, or events into politically loaded categories.
In those cases, Poliscope does not start only by asking whether a single anecdote is true. It also asks whether the category itself is being used responsibly: whether the conduct is clear, comparable cases are treated by the same standard, the evidence fits the category, and the example is part of a meaningful pattern.
If those tests fail, the finding may not be a debunk of one tiny anecdote. It may be something more useful: semantic inflation, selective standard-setting, or discriminatory category-switching.
Why PEM Is Proprietary
PEM is Poliscope's methodology layer. It is the intellectual structure behind how the software keeps claims, receipts, narratives, networks, signals, assessments, and updates from collapsing into one vague impression.
The public principle is straightforward: contested information should be tracked as an evolving evidence record, not as a pile of links or a one-off verdict.
The implementation is proprietary. Poliscope does not publish its internal scoring logic, source-risk tables, graph heuristics, model workflows, taxonomy thresholds, or sensitive operational examples.