N1 2025-07 — audit report

66 questions (問題1–13, the non-listening half). Generated from the OCR and correction audit trails.

OCR confidence

Every section was transcribed twice independently (a full-page pass and a double-resolution half-page pass) and reconciled. Of 13 sections, 0 agreed on all content between the two passes (reconciled automatically, differing only in transcriber notes) and 13 had at least one content difference resolved by a third adjudication pass that re-read the page images. 0 leaf-level differences were examined in total.

Residual OCR risks (13):

• 問題1: transcribed from text witnesses, not the booklet image; see the section-specific audit notes
• 問題2: transcribed from text witnesses, not the booklet image; see the section-specific audit notes
• 問題3: transcribed from text witnesses, not the booklet image; see the section-specific audit notes
• 問題4: transcribed from text witnesses, not the booklet image; see the section-specific audit notes
• 問題5: transcribed from text witnesses, not the booklet image; see the section-specific audit notes
• 問題6: transcribed from text witnesses, not the booklet image; see the section-specific audit notes
• 問題7: transcribed from text witnesses, not the booklet image; see the section-specific audit notes
• 問題8: transcribed from text witnesses; see the section-specific audit notes
• 問題9: transcribed from text witnesses; see the section-specific audit notes
• 問題10: transcribed from text witnesses; see the section-specific audit notes
• 問題11: transcribed from text witnesses; see the section-specific audit notes
• 問題12: transcribed from text witnesses; see the section-specific audit notes
• 問題13: transcribed from text witnesses; see the section-specific audit notes

Source-text quirks the OCR transcribed verbatim (printed typos/recall artifacts; the ones judged to be retype errors were corrected in the layer below) — 15 noted at OCR time:

• 問題1: Q6 answers 2 and 4 are identical in the HTML ('きゅうじゅう'); this is likely a typo on the jlpt247 site — answer 4 may have been intended as 'きゅうちゅう' or similar. Reproduced as-is from source.
• 問題7: The passage text is heavily garbled in the jlpt247 source (apparent OCR/transcription errors from a handwritten reconstruction). Specific garbles noted:
• 問題7: - '誰かも嫌われないで注目されることはめれることや賛そうに見せること?' is heavily garbled.
• 問題7: - '就芯 の仕事をプレーすること?' is garbled; 'プレー' likely should be 'アピール' or 'こなす'.
• 問題7: - 'それだけに考えて' is likely garbled.
• 問題7: - 'それにち良い取り 組のなかに信が加わります' is garbled.
• 問題7: - '私たちを回鍛するために' is garbled; '回鍛' is not a standard word.
• 問題8: Passage (1): jlpt247 garbled the middle clause 'を見せらし、社会は見せようがな' — reconstructed as '…を見れば、社会は見える。社会は見せようがないが、…' for sense. FLAGGED: exact original wording uncertain.
• 問題8: Passage (4): source read '吟味する経験積み重ねると' (missing を) — corrected to '吟味する経験を積み重ねると'. The final sentence's polite/plain register ('…と感知するものである') is reproduced as in source.
• 問題9: q54 (内部 question-62): the jlpt247 source ONLY transcribed option 1; options 2, 3, 4 appear as placeholder '～' in the capture, so they are recorded as empty strings. FLAGGED: q54 options 2-4 are MISSING in the only available source and need to be recovered from another witness.
• 問題9: Passage (1) is heavily garbled in places; reconstructed for sense with the following fixes (FLAGGED as uncertain): '持ようとすれば'→'持とうとすれば'; '変えるさりあげにはならない'→'変えるきっかけにはならない'; the third paragraph ('現問 したとか…獲得したものではないれば') is badly mangled and was reconstructed as '…問題に直面したとか、あるいは実際にそうした領域に足を踏み入れたなどして獲得したものでなければ'; 'これからは求められ れるのではないか'→'これからは求められるのではないか'; '愛付け'→'裏付け'; '変英させて'→'変換させて'.
• 問題9: Passage (2): '混ざり合うや'(注1) — trailing や appears to be an OCR stray and was dropped; (注5) ふつふつと was present in the gloss list but missing its inline mark in the body — inserted after 'ふつふつと'.
• 問題9: Passage (3): 'それを客観的に計る基準に…「好きか、嫌いか」に判断することができる' was garbled — reconstructed as '…「好きか、嫌いか」がある。' (the 'に判断することができる' fragment looks like a duplicated/misplaced clause).
• 問題10: Minor source cleanups (low risk): '楽しんでもらうことはどうすればいいのだろうか'→'楽しんでもらうにはどうすればいいのだろうか'; '作品は「見る」きっかけ'→'作品を「見る」きっかけ'; '巨匠 (4) ○○'→'巨匠○○(注4)' (the (注4) mark was misplaced/garbled in the source); stray trailing 'かい' furigana fragment after '甲斐がある' dropped; '教えるのだけが教育'→'教えるだけが教育'.
• 問題13: FLAGGED — q65 is INCOMPLETE in the only source: the four answer options give just the team NAMES (大空グループ / もみじの会 / 玉北文化の家 / 玉北市歴史クラブ). On the real paper, q65 presents a separate boxed table describing each team's planned 講座 (内容・講師・実施期間・回数・定員・対象者), which the jlpt247 capture did NOT include. That descriptive table is needed to actually answer q65 and is missing here.

Corrections applied (9)

Evidence-based reconstructions of the original exam text. This section is the canonical correction log for the sitting.

• 問3 q14 question: 「<u><u>着手</u></u>していない」 → 「<u><u>着手して</u></u>いない」
  • All four choices are て-forms that combine with the following いない, including the correct 始めて. The tested span is 着手して, leaving いない outside.
  • evidence: structural: 問題3 choices replace the underlined span.
• 問3 q16 question: 「<u><u>破格</u></u>の待遇」 → 「<u><u>破格の</u></u>待遇」
  • The choices are complete adnominal forms, including the correct 特別によい. Leaving の outside the tested span produces 特別によいの待遇; the full target is 破格の.
  • evidence: structural: 問題3 choices replace the underlined span.
• 問1 q5 answers[3]: 「きばく」 → 「さばく」
  • 問題1 reading option 4 (the correct answer): the underlined word is 裁く, read さばく. Our jlpt247 recall capture mis-transcribed it as きばく — not a valid reading of 裁く (さ mis-keyed as き). The independent partial booklet scan prints さばく at position 4. The answer index (3) is unchanged.
  • evidence: input/N1_2025-07/scribd-914932735-exam-scan-partial.pdf (page 1, 問題1 q5)
• 問7 q41 passage: 「」とさわります。」 → 「」と言われます。」
  • Recall garble: 「…教えてください」とさわります is nonsensical; the elder-announcer is being asked, i.e. と言われます. Single-source recall reconstruction of the jlpt247 capture (no exam PDF / second witness exists); a clearly-garbled non-word recovered to its contextually-unambiguous form. Per Domenic's human judgement (OCR-confidence audit notes), implemented 2026-06-17.
  • evidence: Linguistic reconstruction (no witness).
• 問7 q41 passage: 「誰かも嫌われないで注目されることはめれることや賛そうに見せること」 → 「誰からも嫌われないで注目されることや、ほめられることや、偉そうに見せること」
  • Recall garble in a parallel list of rejected purposes: 誰かも→誰からも, はめれる→（や）ほめられる, 賛そう→偉そう. Single-source recall reconstruction of the jlpt247 capture (no exam PDF / second witness exists); a clearly-garbled non-word recovered to its contextually-unambiguous form. Per Domenic's human judgement (OCR-confidence audit notes), implemented 2026-06-17.
  • evidence: Linguistic reconstruction (no witness).
• 問7 q41 passage: 「私はそうたも違うと思います」 → 「私はどちらも違うと思います」
  • Recall garble: そうたも is a non-word; rejecting the two preceding options gives どちらも違う. Single-source recall reconstruction of the jlpt247 capture (no exam PDF / second witness exists); a clearly-garbled non-word recovered to its contextually-unambiguous form. Per Domenic's human judgement (OCR-confidence audit notes), implemented 2026-06-17.
  • evidence: Linguistic reconstruction (no witness).
• 問7 q41 passage: 「それにち良い取り 組のなかに信が加わります」 → 「それにより良い取り組みのなかに信頼が加わります」
  • Recall garble: それにち→それにより, 取り組→取り組み, 信→信頼. Single-source recall reconstruction of the jlpt247 capture (no exam PDF / second witness exists); a clearly-garbled non-word recovered to its contextually-unambiguous form. Per Domenic's human judgement (OCR-confidence audit notes), implemented 2026-06-17.
  • evidence: Linguistic reconstruction (no witness).
• 問1 q6 answers[3]: 「きゅうじゅう」 → 「きゅうちゅう」
  • 問題1 reading question for 胸中 (correct reading きょうちゅう, index 2). The four options are a systematic きょう/きゅう × じゅう/ちゅう distractor set: きょうじゅう(1)/きゅうじゅう(2)/きょうちゅう(3,correct)/きゅうちゅう(4). Option 4 was retyped identical to option 2 (きゅうじゅう), a duplicate; the missing fourth combination is きゅうちゅう. The answer index is unchanged.
  • evidence: jlpt247 witness shows options 2 and 4 both as きゅうじゅう (a duplicate non-distractor, impossible on a real reading item); systematic-distractor reconstruction of the only missing mora-combination. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問6 q40 answers[3]: 「失われうつつある」 → 「失われつつある」
  • Recall garble: 失われ (失われる, passive 'be lost') + つつある ('is in the process of') is the only valid parse — 'limited precious resources are being lost'. 'うつつある' is a non-word; the う is a stray duplicate of the preceding kana. Flagged in the OCR quirk notes as 'likely つつある'.
  • evidence: grammar: つつある attaches to a verb stem (失われ); 'うつつある' is not a word — clear retype artifact. the OCR quirk notes independently flags it; jlpt247 witness reproduces 失われうつつある as-is. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]

Flagged unresolved text issues (2)

Ambiguous source-text issues left unchanged rather than guessed.

• 問7 q41 passage — recall-garble-unrecoverable
  • The 問題7 passage (single-source jlpt247 recall, no exam PDF) contains several garbles whose original wording cannot be recovered without a second witness, left as-is rather than guessed: 「噛一の冗舌が上手になること」, 「就芯の仕事をプレーすること」, 「私たちを回鍛するために」, and 「それだけに考えて言葉に対して予期せぬ反応」. Clearer garbles in the same passage were corrected (see the correction log above).
• 問9 q45 passage — source-divergence-ambiguous
  • 問9(1) matched to 平尾誠二『求心力』(Google Books). Our recall text diverges in minor ways — e.g. 「その程度の『知識』は、持とうとすれば誰でも得ることができる」vs source 「…得ようとすれば誰もが…」, and 「以前は」vs「以前なら」. Left unchanged: as likely intentional exam edits (e.g. avoiding 得…得 repetition) as recall errors, and no exam booklet exists to adjudicate. Source recorded in docs/passage-sources.md.