Advanced Note Taking in Tana with QCE!

QCE—Question, Claim, Evidence—turns note-taking into a structured reasoning system that makes it easy to synthesize what research says while exposing gaps in understanding. Instead of collecting notes as a pile, QCE organizes thinking around specific questions, the competing answers (claims) that literature proposes, and the supporting or opposing evidence for each claim. The payoff is practical: every time a claim gets tagged with evidence, the system can automatically aggregate what’s known for and against it, and the resulting “synthesis” view makes it clear what still lacks coverage.

The workflow starts with questions written explicitly in Tana, such as “Who came up with the idea of a research-focused university model?” For each question, the system expects multiple possible answers. In QCE terms, those possible answers become “claims.” Claims are treated as first-class objects in Tana via a “claim” super tag, which is configured to link each claim back to the exact quote it was derived from. That means a claim like “Wilhelm von Humboldt is responsible for introducing the model of the research university” doesn’t float free; it carries provenance through a linked quote and source.

Evidence is then added through an “evidence” super tag. Evidence entries can support or oppose a claim, and the configuration ensures that evidence automatically accumulates inside the relevant claim. Concretely, the claim super tag includes live searches that pull in all supporting evidence and all opposing evidence associated with that claim. As soon as new evidence nodes are tagged, the claim’s evidence lists update immediately—so the researcher can see, in one place, what the literature supports and what it contradicts.

Once claims are wired to questions, the system scales up. A “question” super tag is set up with a topic field and a synthesis area. The synthesis area uses live searches to collect all claims that inform a given question, plus live searches that track sources through multiple reading stages (e.g., “to process,” “to integrate,” and “integrated”). This staged source pipeline matters because large books often answer multiple questions; a global “read/unread” flag can’t capture whether a source has been revisited for a new angle. By tagging sources with the specific question context, the workspace keeps the reading workflow aligned with the reasoning workflow.

An example centers on the history of the research university model. Quotes attributed to Wilhelm von Humboldt are used to create claims and evidence, and the synthesis view then shows which claims connect to the question, which sources are currently being processed for that question, and what evidence exists for or against each claim. The system also makes missing information visible: if a question’s synthesis contains claims but no sources to process, it signals a gap—such as an unanswered “why” behind the model’s invention.

Overall, QCE in Tana is presented as a way to build a navigable knowledge graph for any topic: questions anchor the structure, claims capture competing answers with citations, evidence accumulates automatically, and synthesis views reveal both what’s supported and what remains uncertain.