Method guide · Progress through phases
Evidence Turnstile
Identify whether a claim rests on observation, experience, or inference before deciding how to use it.
- People
- 4–20
- Base procedure time
- 20 minutes
- Recording
- Shared record
Purpose: Explore through dialogue; Make a decision.
English guide
Text acceptance is not approval of this rendered page, its media, or publication.
How this method works
A conversation in which people identify the basis of their statements as observation, experience, or inference. Use it when direct observations and things heard from others become mixed, inferences are treated as facts, or meetings cannot identify what is missing for a decision. Before making a claim, participants pass one evidence card through the process. Recipients check the type and how it can be used for the decision. Types are not ranked as better or worse; reasons for putting an item on hold are made explicit, making it harder for experienced voices alone to determine the conclusion. The result is an evidence board separating what can be decided now from what needs further confirmation. Do not use it where trust has broken down and speakers’ safety cannot be protected.
Phase schedule
Minutes are cumulative from the start of the base plan, not a live timer. Read scaling for any additional time. Phase controls do not advance automatically.
Use the arrow keys, Home or End while these controls have focus. The complete phase schedule remains below, including without JavaScript.
- Set the issue0–2 minutes from the start (2 minutes)
Action: Set. Participation: Groups of 3–5.
Recording not separately specified for this phase
- Show the evidence2–8 minutes from the start (6 minutes)
Action: Show. Participation: Groups of 3–5.
Individual record
- Check the type8–12 minutes from the start (4 minutes)
Action: Check. Participation: Groups of 3–5.
Shared record
- Separate pending items12–16 minutes from the start (4 minutes)
Action: Separate. Participation: Groups of 3–5.
Shared record
- Choose what comes next16–20 minutes from the start (4 minutes)
Action: Choose. Participation: Groups of 3–5.
Shared record
Facilitation steps
- For 2 minutes, each group writes one issue it wants to decide on an issue card and explains: ‘Evidence types are not a ranking of correctness.’
- For 6 minutes, each person in each group shows one observation, experience, or inference card before stating a claim.
- For 4 minutes, recipients in each group read the evidence type aloud and write one sentence on the display surface about what that evidence allows them to decide now.
- For 4 minutes, put proposals without a type in each group’s pending column and write one checking action and responsible person for each.
- For 4 minutes, each group chooses one item to decide now and one to check next. Combine only those group results into a whole-group list.
If something changes
If
Three minutes after evidence cards start being placed, inference cards still make up at least half of all cards.
Then
Each person writes one situation they saw or heard, changing only cards that can be changed to observation or experience.
If
During the type-checking phase, the same evidence appears 3 or more times.
Then
Bundle the duplicate cards and retain cards individually only where they add new information to the decision.
If
There are 4 or more pending items and selection is not progressing.
Then
Mark only pending items with checking actions possible within the next 3 days. Record the others and close them.
Preparation
Materials
- 6 evidence cards per person: 2 each in 3 colours for observation, experience, and inference
- One issue card per group, approximately A5 size
- One dark-coloured pen per person
- One three-column display surface per group, labelled observation, experience, and inference
Room and setup
- A semicircle for each group of 3–5, with the display surface one step away; 1–4 groups
- One stand per group with evidence cards separated by type
Adjust for people and time
With fewer than 4 people, differences between types are harder to compare, so use no more than 2 evidence cards per person and address only one issue. With 4–20, divide participants evenly into 1–4 groups of 3–5, with no group above 5. Each group uses one display surface, issue card, and card stand. Run all groups simultaneously. Combine only each group’s chosen item and pending conditions at the whole-group level.
Safety and non-applicable situations
Evidence types are not a ranking of correctness. Keep inference distinct from facts, retain reasons for putting items on hold, and do not run the method where speakers’ safety cannot be protected.
Historical translation and source notes
These original notes are retained verbatim with their fixed source versions. This historical label does not resolve a source-specific concern or establish current acceptance.
- Materials specify three columns for observation, experience, and inference, while a later step uses a pending column. The source does not specify whether that is an additional area; this draft preserves both descriptions without changing the layout.
Manuscript's recorded references
These translate the original manuscript's source and rights claims. They remain separate from any current-page notice below; they do not establish new verification.
- Taiwa no Fumen editorial team: Evidence Turnstile
Original description and illustrations
Diagram, print, and video language
Original Japanese guide, diagram, and print view. Their labels and print text have not been translated; an English print edition is not provided here.
English video-library navigation. Original Japanese audio, subtitles and transcripts remain labelled Japanese resources. Their translations are separate later work. No remote player is embedded here.
The manual English phase controls are a navigation aid, not a live facilitation timer. Interactive diagram motion remains pending.
Editorial provenance
Japanese source status and logs are provenance only. They do not mean this English text or interface was field-tested. Historical fixed-text acceptance and hold records remain specific to their recorded versions; displaying them neither asserts current acceptance nor resolves source-specific restrictions.