N1 2015-12 — audit report

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

OCR confidence

Source: retypeset-digital — quality 3/5: Crisp retypeset digital text (Word-to-PDF rendering, grayscale 1408x1990 pages) — fully legible with no scan noise, watermarks, or skew. However, fidelity to the official booklet is mediocre: question numbers are plain '1、' instead of boxed, no furigana, no section tabs or booklet pagination, and several retyping defects: options merged onto the previous line (q21 opt 3, q22 opt 3, q64 opt 1), inconsistent option spacing/indentation (q30 opt 1), and typos in retyped content (高橋克枢 for 高橋克徳, 謙めて for 諦めて, 優秀なやっと for やつと, サービル for サービス, グレープ for グループ). 問題13 says 右のページ though the info sheet follows on the next page. Answer-key correctness cannot be verified from the document itself.

Every section was transcribed twice independently (a full-page pass and a double-resolution half-page pass) and reconciled. Of 13 sections, 4 agreed on all content between the two passes (reconciled automatically, differing only in transcriber notes) and 9 had at least one content difference resolved by a third adjudication pass that re-read the page images. 61 leaf-level differences were examined in total.

Residual OCR risks (24):

• 問題1: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題2: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題3: Instruction-line blank: the source literally prints "問題 3 の言葉" with no ＿＿＿ glyph. Reconstructing ＿＿＿ is a convention-driven normalization (chosen B), not a literal reading; a strict literalist could prefer A's no-blank form. The downstream answer-key pass and section-map are unaffected since the stem text and options are unambiguous.
• 問題3: A half-width space precedes から in the chosen instruction (matching the printed spacing after the bullet group); this is consistent with B's transcription.
• 問題4: q25 option 1 「にともない」 and q23 option 2 「思い詰めて表情を」 are likely retype typos but are preserved exactly per convention; not corrected.
• 問題4: Furigana absent throughout this section in the print (none added), consistent with the source description.
• 問題5: Instruction separator printed as a raised mid-dot (•) run rather than full-width ・; normalized to ・ per conventions — assumed intended as the standard separator.
• 問題5: q34 'カリキュラム改定' (改定) may be a retype variant of the more usual 改訂; transcribed as printed, not corrected.
• 問題5: q28 option 1 'いっさい' is printed in kana where 一切 might be expected; transcribed as printed.
• 問題6: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題7: The instruction's separator is printed with small bullets (•) in the source; per convention both transcribers normalized to full-width 中黒 (・). If the convention intended to preserve the printed glyph, this would differ — but spec says digits/separators follow the 1・2・3・4 form.
• 問題7: The '育ち続けるのだそう' fragment in paragraph 1 is ungrammatical and almost certainly a retype corruption; kept verbatim per 'transcribe exactly as printed' rule, but flagged.
• 問題8: Citation author surname glyph 高橋克权 renders as the simplified 权 in the source; kept as printed per convention (published author is 高橋克徳).
• 問題8: Instruction line and passage-4 citation use middle-dot separators (•) in the source; both transcribers normalized to full-width ・, consistent with the digit/Latin convention.
• 問題9: 尐 (U+5C10) appears three times (passage 2 once, passage 3 once in 尐数, plus none elsewhere) as a retype variant of 少; transcribed as printed per conventions but is genuinely unusual.
• 問題9: 雚 in 熱帯雚林 (passages 3) is an unusual glyph (likely retype error for 雨/樹林); kept as printed in both transcribers and in final.
• 問題9: Numerous in-passage retype typos in passage 1 (はっいて, やっとしか, 考えるずっと, 謙めて, 負けてします) preserved verbatim; correctAnswerIndex left null throughout.
• 問題10: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題11: The 尐 (U+5C10) vs 少 distinction rests on glyph shape in a re-typeset source; if the source font simply renders 少 narrowly, the intended character could be 少. Adjudication transcribes the glyph as printed (尐) for consistency between the two occurrences, since both render identically.
• 問題11: ブラント品 vs ブランド品 is preserved as printed per convention; it is almost certainly a retype typo but is not corrected.
• 問題12: The passage contains three apparent retype typos in the source (数多く刻の / 心の中定着される / 陽光を浴びて茂し). These are transcribed verbatim per convention and flagged in notes; they are not OCR errors but faithful renderings of the flawed re-typeset source.
• 問題12: Both transcribers agree on the entire passage body and all four questions; the only adjudicated content decision was the （中略） paragraph break, resolved confidently against the double-resolution image.
• 問題13: Exact whitespace in the 年会費 line (3 full-width spaces after the box, half-width before 継続入会) is a best visual match to the wide printed gap, not a determinate count; spacing in justified box text is inherently ambiguous.
• 問題13: Whether the boxed labels should be <b> vs <u> depends on the boxed-heading convention; chose <b> per the 問題13 list-glyph/bold-heading rule since they are framed rectangles, not underlines.

Source-text quirks the OCR transcribed verbatim (printed typos/recall artifacts; the ones judged to be retype errors were corrected in the layer below) — 34 noted at OCR time:

• 問題2: Q9: source reads 「希望者の申し込んでください」 which appears to be a retype typo (expected 「希望者は申し込んでください」 or similar); transcribed exactly as printed.
• 問題2: Q9 source reads 「希望者の申し込んでください」, which appears to be a retype typo for 「希望者は申し込んでください」 (の where は is expected); transcribed exactly as printed.
• 問題3: The 問題3 instruction is printed as "問題 3 の言葉に意味が最も近いものを、1•2•3•4 から一つ選びなさい。": the standard ＿＿＿ underline-reference blank is missing and its slot is occupied by the embedded 問題3 label. Per convention, the 問題N label is stripped and the instruction-line blank is normalized to ＿＿＿, yielding ＿＿＿の言葉に意味が最も近いものを、1・2・3・4 から一つ選びなさい。
• 問題4: q23 option 2 reads 思い詰めて表情をしている (missing と after 詰めて would be more natural); retained as printed — possible retype typo.
• 問題4: q25 option 1 reads 国際化にともない (the natural form is に伴って); retained as printed — possible retype typo.
• 問題5: Instruction line: printed separator is a raised mid-dot run (1•2•3•4) and there is a space before から; normalized separators to ・ per conventions and kept the space exactly as printed.
• 問題5: q34: source reads 'カリキュラム改定' (改定, not the more usual 改訂) and 'A大学' (half-width A normalized to ASCII); both transcribed as printed — possible retype variant on 改定.
• 問題5: q28 option 1 'いっさい' printed in kana; transcribed as printed.
• 問題6: q37 answer 1 printed as 「サッカーを続けてこられたのは」; the stem uses 「続けてこられた」 rather than 「続けてこれた」 — kept as printed.
• 問題6: q40 answer 4 printed as 「観賞にたえるような」 (観賞, not 鑑賞) — kept as printed.
• 問題6: q40 option 4 printed as「観賞にたえるような」(観賞, not 鑑賞) — transcribed as printed; possible retype variant.
• 問題6: q38 stem printed as '10 年ぶり' with a space after the ASCII digits; kept text as printed.
• 問題7: Paragraph 1 contains '育ち続けるのだそうお祝いを兼ねて' which reads oddly and is ungrammatical (likely a retype artifact: '育ち続ける' is the stem of the q45 options, apparently mis-inserted, with missing punctuation around 'のだそう'); transcribed exactly as printed.
• 問題7: Option 41-4 is printed in hiragana 'そのゆうじんだった' (友人 written as ゆうじん); transcribed as printed.
• 問題7: Option 45-1 is printed as '育ち続けるのだどうだ' (likely a retype typo for 'のだそうだ'); transcribed exactly as printed.
• 問題7: Passage '○○チャン' uses white circles (U+25CB) as printed, not ideographic zero.
• 問題8: 問題8 passage (1) citation author 高橋克权: printed glyph is the simplified form 权; the published author is 高橋克徳. Likely a retype error but transcribed exactly as printed.
• 問題8: 問題8 passage (2): boxed memo, date 2015年12月1日 right-aligned at top, 以上 right-aligned at bottom; intro line 「以下は、ある会社で回覧された文書である。」 is printed above the box and kept as the passage lead-in. Typo in source: 「弁せてご了承」 (likely intended 「併せて／合わせて」), transcribed as printed.
• 問題8: 問題8 passage (3): printed 「記憶力の身ならず」 is almost certainly a retype typo for 「記憶力のみならず」; transcribed as printed.
• 問題8: 問題8 passages (3)/(4) and citation glyphs verified character-by-character against the double-resolution half-page renders. The instruction line and citation separators are printed as middle dots (1•2•3•4 / 井上俊•船津衛); normalized to full-width ・ per convention.
• 問題9: q50 option 4: 「質も上がられる」 is grammatically odd (likely retype for 「質も上がる」/「上げられる」); transcribed as printed.
• 問題9: Passage (1): suspected retype typos — 「勝負勘はっいてくる」 (likely 「ついてくる」), 「優秀なやっとしか組まない」 (likely 「やつとしか」/「奴としか」), 「考えるずっと映画の質は」 (text seems garbled, likely 「考えるよりずっと」), 「かって」 (likely 「かつて」, appears twice), 「絶対に勝負を謙めてはいけない」 and q51 stem 「謙めてはいけない」 (謙 likely a retype error for 諦), 「結果的に負けてします」 (likely 「負けてしまう」). All transcribed as printed.
• 問題9: q55 options 1 and 4 printed in 「1.」/「4.」 narrow format and physically split: option 4 appears after a blank line; option 2 wording 「欲望を満たすのは必要な時間は短くならないから」 reads oddly (likely 「満たすのに」), transcribed as printed.
• 問題9: Passage (2): 「尐なかった」 uses 尐 (U+5C10), as printed — appears to be a retype variant of 少.
• 問題9: Passage (3): 「熱帯雚林」 uses 雚 (likely a retype error for 雨-radical 樹林/雨林; the OCR glyph resembles 雚), appears twice; transcribed as printed. 「男が大きくなるところか」 likely 「どころか」. q57 option 4 typo 「グレープ」 (should be グループ); q57 options 1/4 and q58 options 1/4 printed in narrow 「1.」/「4.」 format. All transcribed as printed.
• 問題10: Suspected retype typo in passage: 「目的も手段も違うものなもだとしか」— 「ものなもだ」 appears garbled (likely intended 「ものなのだ」); transcribed exactly as printed.
• 問題10: （中略）appears inline at the end of the 天文学の先生 quotation paragraph (after 「…教えてくれました。」), not on its own printed line; kept inline as printed.
• 問題11: Passage A prints 「ブラント品」 (kana ト); passage B writes 「高級ブランド品」. The former is very likely a retype typo for ブランド品 but is transcribed as printed.
• 問題12: 注1 gloss reads 「あまたの：数多く刻の。」 in the source; '刻の' appears to be a retype typo (likely intended 数多くの), transcribed as printed.
• 問題12: Passage line 「見る者の心の中定着される山のイメージ」 appears to be missing a particle (likely 心の中に定着), transcribed as printed.
• 問題12: Passage 「陽光を浴びて茂し」 looks like a retype typo (likely 茂り), transcribed as printed.
• 問題12: 注5 gloss uses a half-width colon (実相:本当の姿。) while notes 1-4 use full-width 「：」; transcribed as printed.
• 問題13: q69 stem prints “サービル” (a retype typo for “サービス”); transcribed as printed.
• 問題13: Passage bullet “館内のショップで10%の割引が受けられます。ただし、書籍はの除きます” contains an apparent retype typo (“書籍はの除きます”, likely intended “書籍は除きます”); transcribed as printed.

Corrections applied (36)

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

• 問1 q6 question: 「練習に<u>励ん</u>でいきます」 → 「練習に<u>励んで</u>いきます」
  • The printed underline runs under 励んで (including で); our span drops the で. The answer options (はげんで etc.) all include the で, confirming the reading target is 励んで.
  • evidence: underline-boundary-vs-image
• 問2 q9 question: 「希望者の申し込んでください」 → 「希望者のみ申し込んでください」
  • 希望者の申し込んで is ungrammatical; original reads 希望者のみ申し込んで (only applicants apply). The quirks file's は-hypothesis is wrong.
  • evidence: jlpt-freq witness reads 希望者のみ申し込んでください [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問4 q23 answers[1]: 「思い詰めて</u>表情をしている」 → 「思い詰めた</u>表情をしている」
  • Correct-answer option must use the word naturally; 思い詰めた表情 is the standard noun-modifying form, 思い詰めて表情 is ungrammatical.
  • evidence: jlpt-freq witness reads 思い詰めた表情をしている [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問6 q40 answers[3]: 「観賞にたえる」 → 「鑑賞にたえる」
  • The idiom 鑑賞にたえる (worth artistic appreciation) describes an animated film; 観賞 (viewing flowers/nature) is wrong.
  • evidence: jlpt-freq witness reads 鑑賞にたえるような [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問7 q45 passage: 「育ち続けるのだそうお祝いを兼ねて」 → 「育ち続けるのだそうだ。お祝いを兼ねて」
  • Drops sentence-final だ and the period, fusing two sentences; witness shows a clean break.
  • evidence: jlpt-freq witness reads 育ち続けるのだそうだ。お祝いを兼ねてアパートを訪ね [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問7 q45 answers[0]: 「育ち続けるのだどうだ」 → 「育ち続けるのだそうだ」
  • のだどうだ is a non-word; どうだ is a retype of そうだ.
  • evidence: jlpt-freq witness lists option 1 as 育ち続けるのだそうだ [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問8 q46 passage: 「高橋克权」 → 「高橋克徳」
  • The author of the cited title is 高橋克徳; 权 is a corruption of 徳.
  • evidence: Google Books HIT: 「上司がさっぱりわかってくれない」と思っているあなたへ 高橋克德 (Katsunori Takahashi) [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問8 q47 passage: 「弁せてご了承」 → 「併せてご了承」
  • 弁せて is a non-word; intended phrase is 併せてご了承 (please also note).
  • evidence: jlpt-freq witness reads 併せてご了承ください [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問8 q48 passage: 「記憶力の身ならず」 → 「記憶力のみならず」
  • 記憶力の身ならず is a non-word; the following 計算力とか、情報整理力とか list requires のみならず (not only memory).
  • evidence: grammar: のみならず; 身/み homophone-OCR confusion [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q50 passage: 「勝負勘はっいてくる」 → 「勝負勘がついてくる」
  • はっいて is a non-word (stray small っ); idiom 勘がつく (intuition develops).
  • evidence: jlpt-freq witness reads 勝負期はっいてくる, confirming ついて; idiom 勘がつく [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q50 passage: 「優秀なやっとしか」 → 「優秀なやつとしか」
  • やっとしか is a non-word; intended is やつとしか (only with capable guys).
  • evidence: jlpt-freq witness reads 優秀なやつとしか組まない [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q50 answers[3]: 「質も上がられる」 → 「質も上げられる」
  • Correct-answer option must be grammatical; 上がる is intransitive and cannot take 上がられる; correct is 上げられる.
  • evidence: jlpt-freq witness reads 質も上げられる [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q51 passage: 「絶対に勝負を謙めてはいけない」 → 「絶対に勝負を諦めてはいけない」
  • 謙める is not the intended word; 諦める (give up) is meant.
  • evidence: jlpt-freq witness reads 絶対に勝負を諦めてはいけない [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q51 question: 「絶対に勝負を謙めてはいけない」 → 「絶対に勝負を諦めてはいけない」
  • Same typo in the q51 stem quoting the passage; 謙→諦.
  • evidence: jlpt-freq witness reads 絶対に勝負を諦めてはいけないとあるが [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q50 passage: 「結果的に負けてします」 → 「結果的に負けてしまう」
  • 負けてします is a non-word; intended is 負けてしまう (end up losing).
  • evidence: jlpt-freq witness reads 結果的に負けてしまう [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q55 passage: 「できる事柄が尐なかったころ」 → 「できる事柄が少なかったころ」
  • 尐 (U+5C10) is a non-standard glyph for the intended exam character 少.
  • evidence: jlpt-freq witness reads できる事柄が少なかったころ [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q55 answers[1]: 「欲望を満たすのは必要な時間は短くならない」 → 「欲望を満たすのに必要な時間は短くならない」
  • のは is wrong; parallels the passage's 欲望を満たすのに必要な時間.
  • evidence: jlpt-freq witness q55 option 2 reads 欲望を満たすのに必要な時間は短くならないから [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q57 passage: 「力のある男たちが尐数のグループ」 → 「力のある男たちが少数のグループ」
  • 尐 (U+5C10) is a non-standard glyph for 少 (少数).
  • evidence: jlpt-freq witness reads 体力のある男たちが少数のグループ [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q56 passage: 「熱帯雚林を出よう」 → 「熱帯雨林を出よう」
  • 雚 is an OCR corruption; the word is 熱帯雨林 (tropical rainforest).
  • evidence: jlpt-freq witness reads 熱帯雨林を出ようとしていた [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q56 passage: 「熱帯雚林の外では」 → 「熱帯雨林の外では」
  • 雚 is an OCR corruption; the word is 熱帯雨林.
  • evidence: jlpt-freq witness reads 熱帯雨林の外では [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q58 passage: 「男が大きくなるところか」 → 「男が大きくなるどころか」
  • ところか is wrong here; the expression is どころか (far from).
  • evidence: jlpt-freq witness reads 男が大きくなるどころか [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q57 answers[3]: 「体力のある者のグレープ」 → 「体力のある者のグループ」
  • グレープ (grape) is a non-word; intended is グループ (group). Correct-answer option.
  • evidence: jlpt-freq witness q57 option 4 reads 体力のある者のグループ [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問10 q59 passage: 「目的も手段も違うものなもだとしか」 → 「目的も手段も違うものなのだとしか」
  • ものなもだ is garbled; intended is ものなのだ.
  • evidence: jlpt-freq witness reads 違うものなのだとしか言いようがない [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問11 q63 passageA: 「ブラント品」 → 「ブランド品」
  • ブラント is a katakana typo for ブランド; passage B writes 高級ブランド品.
  • evidence: jlpt-freq witness reads ブランド品; passage B uses ブランド [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問11 q63 passageA: 「収入が尐なく」 → 「収入が少なく」
  • 尐 (U+5C10) is a non-standard glyph for 少.
  • evidence: jlpt-freq witness uses 少; OCR transcriber A read 収入が少なく [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問11 q64 answers[3]: 「魅力的なものが尐ない」 → 「魅力的なものが少ない」
  • 尐 (U+5C10) is a non-standard glyph for 少.
  • evidence: jlpt-freq witness reads 魅力的なものが少ない; OCR transcriber B read 少ない [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問13 q69 question: 「サービル」 → 「サービス」
  • サービル is a non-word typo for サービス; 会員サービス appears correctly elsewhere in the same passage.
  • evidence: jlpt-freq witness reads サービスはどれか [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問13 q69 passage: 「書籍はの除きます」 → 「書籍は除きます」
  • はの is a stray-particle artifact; intended is 書籍は除きます (books excluded).
  • evidence: jlpt-freq witness reads 書籍は除きます [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q50 passage: 「僕一人で何も考えるずっと映画の質は高くなるのだ」 → 「僕一人で何も考えずとも映画の質は高くなるのだ」
  • 「考えるずっと」is a non-grammatical garble. The clause means 'even without my thinking, the film's quality rises (because I only team with excellent people)'. The intended form is 考えずとも (= even without thinking); ずっと is a retype corruption of ずとも.
  • evidence: Category typo, confidence medium. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問9 q57 answers[1]: 「共に食物を探しように」 → 「共に食物を探し歩くように」
  • 「探しように」is a non-word; 〜ようになった attaches to a dictionary form, so the verb must be 探す, not the 連用形 探し. Minimal grammatical fix.
  • evidence: Category typo, confidence medium. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問12 q65 passage: 「見る者の心の中定着される山のイメージ」 → 「見る者の心の中に定着される山のイメージ」
  • Particle に dropped: 定着する takes a location with に (心の中に定着される). Parallel to the preceding 「心に刻印される」.
  • evidence: Category dropped-char, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問12 q65 passage: 「植物たちは陽光を浴びて茂し、渓流は」 → 「植物たちは陽光を浴びて茂り、渓流は」
  • 「茂し」is not a valid verb form. In this list of 連用中止 forms (舞い…茂り…削り), the verb 茂る yields 茂り. 茂し is a retype error.
  • evidence: Category typo, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問12 q65 passage: 「（注1）あまたの：数多く刻の。」 → 「（注1）あまたの：数多くの。」
  • The gloss for あまた should read 数多くの ('numerous'). The stray 刻 (which appears in nearby gloss 注3 刻印する) is a retype intrusion; 数多く刻の is not Japanese.
  • evidence: Category typo, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問9 q0 passage: 「かって負けた」 → 「かつて負けた」
  • かって is a kana-confusion typo for かつて (formerly); the passage context (past defeat) requires かつて.
  • evidence: Category homophone, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問9 q0 passage: 「（注3）に映」 → 「に（注3）映」
  • （注N）note marker repositioned to follow its glossed term (the retype placed the marker before the word).
  • evidence: Marker-only move; the marker-stripped text is byte-identical. [note-marker relocation pass, 2026-06-20]
• 問11 q0 passageB: 「（注1）すると」 → 「する（注1）と」
  • （注N）note marker repositioned to follow its glossed term (the retype placed the marker before the word).
  • evidence: Marker-only move; the marker-stripped text is byte-identical. [note-marker relocation pass, 2026-06-20]