Japanese name & reading ledger

Three decisions belong in separate fields

Written form, pronunciation and creative interpretation are different kinds of information. A kanji gloss does not establish a person-name reading, and a pronunciation chosen for a fictional character does not become an attested cultural fact because it looks plausible. This ledger keeps the fields apart. Missing kana remains unknown. No dictionary lookup or reading prediction runs behind the scenes.

Choose an honest reading status

Approved means the creator has supplied and accepted the reading. Invented means it is an explicit fictional choice. Attested means the user is making a source-based claim and must provide a reading and a source URL. The tool does not verify that the URL supports the claim. Unknown preserves the absence of evidence. These labels represent different provenance, not a ladder on which an invented choice must be disguised as a real one.

What normalization does and does not do

For collision screening, the tool normalizes width and compares katakana with corresponding hiragana forms. Original display strings remain intact. A matching comparison form flags that readers or performers may encounter the same supplied reading under two written names. It does not merge people, approve a spelling variant, establish historical equivalence or judge whether a name sounds natural.

Fictional example: the library card

A comic shows a library card before its owner introduces themselves. The author withholds the reading so that a narrator’s assumption can later be corrected. If the later reading is invented, the production ledger labels it invented. If the clue relies on a documented alternative reading, that particular claim needs a suitable source. A general kanji definition would not establish the alternative.

Make the clue fair in the medium

Ruby can reveal a reading that an image alone cannot. A localization may need a different clue because romanized text exposes information earlier. An audiobook has no silent written ambiguity to preserve. Ask what the reader is able to perceive at each point. The name should produce a specific change in interpretation, not depend on the audience accepting an unsupported assertion about Japanese.

Evidence and review remain necessary

A source entry needs a locator and a clear claim: what exactly does it substantiate? A general institutional rule about written names and recorded readings does not validate your candidate. Font coverage and glyph fallbacks are separate production questions. Keep the original form, the approved reading, the evidence state and any editorial substitution available to the translator and letterer. Native-language review is especially important when the setting presents the name as realistic rather than deliberately fictional.

Transfer a source record without inventing a reading

The separate library searches the imported JMnedict snapshot by written prefix or supplied kana. A spelling may have several source entries and readings; their IDs stay separate. Mixed source classifications, such as a form used as both a surname and a place name, remain visible. Using a record appends the source pairing to your ledger without removing authored rows. Dictionary inclusion does not establish contemporary usage, a birth cohort, registration eligibility, etymology or suitability for your character. Preserve the EDRDG attribution and applicable licence when handing on derived records.

Build a record whose claims can be inspected

Use one stable ID for each fictional person. The written-form field preserves the text the creator selected, while the kana field records a supplied reading. The status explains where that reading came from. The source URL is not a general badge of authenticity: a useful reference must support the exact claim being made. A page about a character’s general meaning does not necessarily support a particular name reading.

The evidence table adds a source locator, claim scope, creative association and reuse declaration. A locator might identify an entry or a specific passage. The scope should say whether the source supports a spelling, a reading or some other limited point. The creative-association field belongs to the author’s interpretation. It must not migrate into a statement that the name historically or universally means what the author wants it to suggest.

The workbench does not fetch the URL or independently authenticate an entry. Evidence remains unverified unless an actual reviewer checks it outside the tool. This restriction prevents a plausible-looking link from becoming a false assurance. It also keeps private manuscript details from being silently submitted to a dictionary or an AI service. The local report can tell you what still needs checking without pretending to have completed that work.

Work through two forms and one reading

The initial example intentionally gives two fictional written forms the same supplied reading, one entered in hiragana and the other in katakana. Both are labeled invented. The normalization screen can find that comparison without claiming that either is an attested personal name. Original spelling and kana remain unchanged in the output.

Change the reading of only one record and run again. The collision changes because the evidence you supplied changed. Changing a creative gloss should not alter the reading. Leaving kana empty should produce unknown, not a pronunciation guessed from individual characters. These are separate tests of the data model, not questions of whether a name feels convincing.

If you select attested, provide both kana and a source URL. Then add the source locator and scope. Missing information should remain a review question. Do not switch the status to approved simply to suppress a missing-reference error if you still intend to publish an external factual claim. Author approval establishes what your fictional character is called; it does not establish a historical or linguistic assertion.

Original fictional scene: the card and the introduction

The narrator first saw the name on a library card. He put the card beside the returned book and used the reading he assumed was right.

The visitor did not answer. He repeated it more clearly, thinking the rain outside had swallowed the first attempt.

“That is not how I say it,” she said.

He looked back at the card. The print had not changed. He wanted it to explain the mistake for him.

“Would you write the reading?”

She wrote it on a separate slip and placed it across the lower edge. The librarian’s stamp stayed visible. Her correction did not make the earlier assumption a fact about the card; it showed that the assumption had been his.

Later, when another character found the borrowed book, the narrator almost repeated the old reading while describing its owner. He stopped and used the form she had given him. The pause revealed that he remembered the correction. It did not reveal a secret ancestry or an official rule he had somehow learned from the characters alone.

At the end of the chapter, the comic printed the supplied reading beside the name. The production ledger recorded it as an invented choice. The reader could now connect the spoken introduction to the earlier image without being told that a dictionary had established something it had not.

This is a fictional scene about an assumption and a correction. The example does not provide an attested reading for its invented name. If a different story depends on a documented alternative reading, that specific alternative needs evidence of its own.

Three versions of the same naming decision

In a realistic version, the creator chooses an attested form and reading with suitable evidence, then asks a qualified language reviewer to assess the context. The story can use the narrator’s error, but it should not imply that the incorrect reading is universally obvious or that every speaker would make the same assumption.

In an explicitly invented version, the author assigns the reading as part of the fictional world. The ledger labels that choice and the article describes it as creative construction. The characters’ reactions must be supported by the world’s stated conventions. There is no need to disguise invention as documentary fact to give the scene emotional weight.

In a purely visual version, the clue relies on an ink mark, an obscured character or a prop alteration. The reading may be irrelevant. The proof then belongs partly in the lettering and asset-continuity workflow. Treating every graphical detail as a linguistic mystery would add unsupported complexity without helping the reader solve the actual problem.

Reading support changes what the audience knows

Ruby, inline kana, a spoken introduction and a romanized credit reveal different information at different times. A silent image can leave pronunciation unspecified; an audio performance cannot preserve that exact silence once the name is spoken. A translated edition may expose a connection earlier because its ordinary display form includes pronunciation information.

Plan the sequence rather than assuming the same clue survives every medium. Which form appears first? When can the reader connect it to a speaker? Does a glossary reveal the reading before the intended scene? Keep the public introduction separate from the private production record, where artists and performers may need the information in advance.

A withheld reading is not automatically an effective mystery. If the eventual explanation requires knowledge the story never supplied, it may feel arbitrary. Make the consequence observable: a corrected address, a misdirected record or a changed inference. Then explain why another name or an earlier reading note would change that consequence.

Handoff without erasing provenance

The letterer-and-translator export pairs the stable ID, written form, supplied reading, status and evidence fields. Plain-text output retains the status even when ruby markup is unavailable. A private editorial handoff can include unverified claims as questions, clearly labeled; it should not be republished as a verified dictionary.

The corpus export excludes source-dependent entries whose reuse declaration is not cleared. A user declaration remains a declaration, not independent license clearance. Source files and intended reuse need their own review before redistribution. Do not assume that a publicly readable entry can be copied into a commercial data product without checking the actual conditions.

Finally, check glyphs separately from readings. A font fallback might preserve a readable shape while losing a deliberate prop distinction. The original written form, approved display substitution and source evidence should remain separate. Correcting a reading must not silently replace the character’s identity, and a graphic normalization must not establish that two similarly written people are the same person.

Does approved mean this is a documented real reading?

No. It means the creator has accepted the supplied reading for the character. Attested is a separate source-based claim and still needs verification of its exact evidence. Invented openly identifies a fictional choice. These statuses should survive exports. They are not marketing grades, and moving between them is not a way to avoid explaining a claim you still intend to present as factual.

Why compare hiragana and katakana without rewriting my record?

A comparison form can identify an equal supplied reading while the original script remains useful to production. Normalization is not an editorial instruction to replace the display. It also does not establish historical glyph equivalence or merge people. Preserve stable IDs and source forms so a collision warning cannot turn into a silent identity change.

Can I export an unverified source entry to a translator?

A private editorial handoff can carry an explicitly unverified question with a locator and scope. That is different from redistributing the material as a verified name corpus. The eligible-corpus export applies the declared reuse filter. A user clearance declaration still needs real review for the intended source and use; it is not a legal conclusion produced by the tool.

What should a creative meaning explanation say?

Describe it as an author association or an invented world gloss. Explain how it affects this character or scene rather than asserting a universal cultural meaning. A dictionary character gloss does not automatically support a full personal-name etymology. If the story needs a real historical assertion, attach evidence for that assertion and preserve its scope and uncertainty.