How Arbitir works
Five things to understand before you run your first analysis.
01
What Arbitir measures
Arbitir does not check whether the facts in a piece of content are correct. Fact-checking requires knowing what is true. Arbitir measures something different: whether the reasoning structure of the argument is sound — regardless of whether the underlying facts are right or wrong. A piece of content can be factually accurate and logically incoherent. It can be factually wrong and well-reasoned. Arbitir scores the reasoning. Verifying the facts is your job.
02
The five circles
Every analysis scores the content across five structural reasoning dimensions.
Missed Clues
Evidence present in the source that contradicts or complicates the argument.
Ignored Other Side
Alternatives that were available and not acknowledged.
Jumped To
Conclusions that outrun the evidence provided.
Untested Assumption
Premises treated as established that were never verified.
Blind Spot
Structural bias in how the argument was framed — including AI policy layer signals in AI-generated content.
03
The grade
A through F. A = reasoning holds under scrutiny. F = the argument fails on multiple structural dimensions. The grade is a composite of the five circles. It is not a measure of factual accuracy, writing quality, or editorial position.
04
The Reasoning Mode
Every analysis returns a Reasoning Mode: Analytical, Experiential, or Integrated. This describes the structural character of the argument — how it was built, not how well.
05
What Arbitir does not analyze
Religious texts. Content depicting harm to minors. Content promoting self-harm. Arbitir applies no political, cultural, or editorial standard. It applies a cognitive one.