N1 2019-07 — audit report

69 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: Born-digital retypeset PDF (text layer present but its ToUnicode cmap is garbled, so extracted text is mojibake): crisp, fully legible type on clean white pages. Marred by gray correction-patch overlays scattered across many pages, tofu/black-box corruption in a few instruction lines (問題2 header p2, 問題5 listening intro p26, one answer-key column header), a badly corrupted embedded scan image for 問題8 passage (2) on p8, and the question-numbering discrepancies noted in 'missing'. Cover note itself warns the file is not a final version.

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. 72 leaf-level differences were examined in total.

Residual OCR risks (27):

• 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: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題4: Q21 option 1 reads '話し合いで' on the page; if the intended exam answer key word were '話し合って' this would be a source typo, but as printed it is 'で' and is transcribed verbatim.
• 問題4: Underline endpoints around '業績' in Q22 option 3 were judged from rendering; small antialiasing on the underline tail is the only ambiguity, but both high-res halves support the span ending at '業績'.
• 問題5: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題6: Instruction-line blank rendering: the printed blank is an underlined run containing a ★; rendered as ＿＿＿★＿＿＿. This is a judgment call not directly covered by the documented '＿＿の言葉' (＿＿＿) convention.
• 問題6: The stray intra-word spaces in Q37/Q40 (and the 'X 医院' gap) are retained per the transcribe-exactly rule but could be normalized differently by a downstream consumer.
• 問題7: The second 【44】 marker in the passage is a printed typo (should be 45); kept as printed 【44】 for fidelity, so the passage carries 41,42,43,44,44 and no literal 45 marker.
• 問題7: 注 parentheses print half-width as (注 1)/(注 2) with a space, but are normalized to full-width （注1）/（注2） per the 注-gloss convention example; deliberate normalization, not verbatim.
• 問題7: The 「自分は、… opening quote has no closing quote in print; transcribed as-is (source typo).
• 問題8: Passage (2) is a badly corrupted vertical-text image with solid black boxes; exact ■ counts and the identity of obscured characters are inherently uncertain. The chosen run lengths (e.g. 最■も, 住い■、そこ■■む, ■■まに, ときに■■■うな, 生■■様, 世界■に■■土■) are best-effort reads of how many character cells are fully blacked out and may be off by one in places.
• 問題8: 最■も: 最 and も are both legible and 最も is the natural word, so the single black box between them may be rendering corruption rather than an obscured character; kept ■ to flag the visible artifact, but 最も (no ■) is a defensible alternative.
• 問題8: The apparent space in body markers like '(注 1)' and gloss '（注 2)' is treated as kerning between full-width 注 and an ASCII digit (per the half-width-digit convention) and omitted, matching both transcribers; if it is a real space, the markers would be '(注 1)' etc.
• 問題9: Bold question-stem phrases that quote the passages are not underlined in this Chinese-prep retypeset; if the official sheet underlined them, the <u> spans are absent here (both transcribers agreed on omission).
• 問題9: Source is a re-typed Word document; passage (1) plausibly contains retype typos but is transcribed exactly as printed (A) per convention.
• 問題10: Gray correction overlays sit over a few passage and answer spans; text beneath was legible in the double-res crops and matches surrounding clean readings, but a tiny chance the overlay hides a one-character correction cannot be fully excluded.
• 問題10: B reported a doubled 「生き残った…」 garble in para 2; the double-res image clearly shows the single clean reading 「生き残ったものがいた。そして、その生物たちは、」, so this is judged a B OCR error, not a source defect.
• 問題11: No 注 gloss text is printed under either passage on pages 18–19, even though passage A carries a (注) reference mark; the gloss may have been dropped by the prep-school retypeset. correctAnswerIndex left null for both questions as required.
• 問題11: The duplicated 陥陥り and the missing-さ 早 are preserved as printed retype artifacts; if a downstream consumer expects clean Japanese it should treat these as known source typos, not transcription errors.
• 問題12: The 2-em dash in paragraph 3 ('なかには――…なかにも――') is rendered ―― (two U+2015) per convention; the printed glyph is a long unbroken rule whose exact code-point/length cannot be distinguished from the image, but ―― is the established normalization.
• 問題12: The 注 markers are printed '(注 N)' (half-width parens + space); normalized to full-width （注N） per established project convention. If the convention were ever to preserve half-width parens, every （注N） would need revisiting.
• 問題12: Several light gray patches (correction overlays) sit over passage and option text; in every checked case the underlying glyphs were legible and used, but a patch could in principle hide a single-character edit not represented in the rendered-under text.
• 問題12: The passage is a Chinese prep-school retype; a few wordings (e.g. 'いってよい自由', 'そこに没入したい（注3）世界') read slightly unnaturally but were transcribed exactly as printed and are confirmed legible, not OCR guesses.
• 問題13: Section A 会場 line spacing: '各回50名' before '◇会場' may be one or two full-width spaces; chose single (matches agreed D-course line). B used two.
• 問題13: Numeric thousands comma '1,000'/'1,800' kept as ASCII comma (both transcribers agreed); printed glyph could be full-width '，'.
• 問題13: 見学先 '旧前岩邸' — 岩 kanji slightly ambiguous in the render.

Source-text quirks the OCR transcribed verbatim (printed typos/recall artifacts; the ones judged to be retype errors were corrected in the layer below) — 30 noted at OCR time:

• 問題1: Question 1 option 1 "もれつ" and option 4 "もうれつ" are distractors as printed; transcribed exactly.
• 問題2: The 問題2 instruction line is partially corrupted/illegible in the source: it reads '（　　）に入れるのに最も上■■■■。' where the expected wording would be '最もよいものを'. The garbled portion (after '最も') is transcribed best-effort as '上' followed by four unreadable boxed glyphs.
• 問題3: The instruction line and 問題3 label render in a corrupted/garbled font on this re-typeset PDF; the instruction text is reconstructed as the standard 問題3 wording (＿＿＿の言葉に意味が最も近いものを、1234から一つ選びなさい。). The point annotation printed after it reads (1*6) i.e. 1 point x 6 questions.
• 問題3: Several option groups (Q18 ばてて/おなかがすいて/実行する etc.) have light gray correction-overlay shading on parts of the text; treated as overlays per instructions, not answer marks, and transcribed as printed.
• 問題4: Q21 option 1: '話し合いで' is correctly printed (A misread it as the typo '話し合いて'); '話し合いで' matches the image.
• 問題4: Q22 option 3: the underline spans 'ほほえましい業績' (both the headword and the following 業績) as printed.
• 問題5: Questions use half-width ? in dialogue (来るんだよね?, 見つかった?, etc.) as printed.
• 問題6: Q37 source prints 'インス タント食品' with a stray space inside the word (retype artifact); transcribed as printed (space kept).
• 問題6: Q40 source prints '生 徒たち' with a stray space inside '生徒' (retype artifact); transcribed as printed (space kept).
• 問題6: Q38 scene label printed as （ホームページーで）with an extra long-vowel mark after ホームページ — transcribed as printed (likely a retype typo).
• 問題7: The last cloze blank in the passage (preceding 「裏側に、真っ暗で人に見せたくない部分を」) is printed as a boxed 44, the same number as the earlier blank, but it is clearly the 45th blank given that questions 41-45 are present and the answer set for 45 fits this slot. Transcribed as 【44】 per the printed box; this is almost certainly a retype typo for 45.
• 問題7: In the paragraph beginning 「それからもう一つは」, the source has an opening 「 on 「自分は、相手に応じて…」 with no matching closing quote before 顔をする — transcribed exactly as printed (suspected source typo).
• 問題8: Passage (1): printed text reads 'ポイントになる だろう。' with an extra half-width space before だろう — transcribed as printed.
• 問題8: Passage (2) is a heavily corrupted embedded vertical-text (縦書き) block: the source render has many solid black boxes obscuring characters. Transcribed left-to-right horizontally; fully-obscured character cells are marked ■, legible characters kept. Best reading of the corrupted runs (from double-resolution crops): 人間の最[■]も…/ 住い[■(=は)]、/ そこ[■■(=に住)]む / [■■(=さま)]まに / 表[■]れる / 世界[■]に[■■]土[■] / ときに[■■■(=このよ?)]うな表現 / 人間の生[■■(=の多)]様な在り方. The ■ spans are inherently uncertain.
• 問題8: Passage (2)'s 注 marker is printed as （注） inline (no number) with FULL-WIDTH parens, unlike the other glosses; transcribed as printed.
• 問題8: Q49 stem uses 述べている as printed.
• 問題9: Passage (2) on p12: paragraph 3 ends mid-sentence at つぎ and paragraph 4 begins with のステップ, an awkward printed line/paragraph break in the source, transcribed as two separate <p> as printed.
• 問題9: Q57 option 1 reads ひとに決められたことの中には… — transcribed as printed.
• 問題10: Gray patches over the tail of 問題61 answers 3 (「…べきである。」) and 4 (「…ものだ。」), plus over some passage spans (e.g. para 2 「また地上の」, para 3 「彼らは我々が住みやすい」), are correction overlays, not answer marks; text underneath is legible and transcribed as printed.
• 問題11: Passage A: 「決めるために必要な情をたくさん持っていたり」 is missing 報 (likely a retype typo for 必要な情報) — transcribed as printed.
• 問題11: Passage B: the source duplicates 陥 across a line break (line ends 「…状況に陥」, next line begins 「陥りやすい」), yielding 「…状況に陥陥りやすい」 — transcribed as printed.
• 問題11: Passage B: 「しかし、早が必要だとはいっても」 is missing the expected さ (likely intended 早さが必要) — transcribed as printed.
• 問題11: Passage B: 「多角的に物事をとらえられようになることが必要だ」 reads oddly (expected とらえられるように); the source breaks the line after とらえられように, which is rendered as continuous text per convention — transcribed as printed.
• 問題11: Q62 answer 4: an extra full-width gap appears before ができる in 「回避すること　ができる」 — preserved as a single half-width space as printed.
• 問題12: 問題12 is a single long passage with printed questions 64-67, matching expectation. The four question stems and all sixteen answer options (p21) are clear and were verified character-by-character against the double-resolution halves; both transcribers had them correct.
• 問題12: Underlining: ①<u>あの自由</u>, ②<u>かわいそうな子どもたち</u>, ③<u>偉大な作品</u> are underlined in BOTH the passage body and the question stems; circled marks ①②③ kept as printed immediately before the underline (B included these in the passage; A omitted the passage-body underlines).
• 問題13: Instruction is printed with stray spaces ('問いに 対する答え' and '1・2・3・4 から'); transcribed as printed. The leading '問題13' label and a trailing '(2*2)' point/retype marker are ignored per spec.
• 問題13: Postal code and phone number use a full-width dash (U+2212): '〒027−3351', '019−160−0224'; kept as printed (not normalized to ASCII '-', since only digits/Latin letters are normalized).
• 問題13: Question 69 stem has two stray spaces as printed: '再募集に申し 込もうと' and '受講料の支払い はどうすれば'.
• 問題13: Answer numbers '1,000円'/'1,800円' are printed with a thousands-separator comma; kept as printed.

Corrections applied (17)

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

• 問3 q17 question: 「<u>不審</u>な点」 → 「<u>不審な</u>点」
  • The choices are complete adnominal forms, including the correct 怪しい. Leaving な outside the tested span produces 怪しいな点; the full target is 不審な.
  • evidence: structural: 問題3 choices replace the underlined span.
• 問3 q19 question: 「<u>まっとう</u>する覚悟」 → 「<u>まっとうする</u>覚悟」
  • All four choices already end in する, so leaving する outside the tested span duplicates the verb. The full target is まっとうする.
  • evidence: structural: 問題3 choices replace the underlined span.
• 問4 q22 answers[2]: 「<u>ほほえましい業績</u>」 → 「<u>ほほえましい</u>業績」
  • Headword ほほえましい; the printed underline covers only ほほえましい and ends before 業績. Our span wrongly swallows the following word 業績.
  • evidence: underline-boundary-vs-image
• 問6 q37 question: 「インス タント食品」 → 「インスタント食品」
  • Stray space inside the word インスタント is a retype artifact; the original prints インスタント食品 as one word.
  • evidence: jlpt-freq witness (n1-2019.7真题) lines 283 and 2053: 「インスタント食品」 with no space. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問6 q38 question: 「（ホームページーで）」 → 「（ホームページで）」
  • Extra long-vowel mark after ホームページ yields the non-word ホームページー; the scene label should read ホームページで.
  • evidence: clear non-word (ホームページ is already complete); audio-script witness writes ホームページから with no trailing ー [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問6 q40 question: 「生 徒たち」 → 「生徒たち」
  • Stray space inside the word 生徒 is a retype artifact.
  • evidence: jlpt-freq witness (n1-2019.7真题) lines 313 and 2104: 「生徒たちに」 with no space. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問7 q45 passage: 「見える。【44】裏側に」 → 「見える。【45】裏側に」
  • The fifth cloze blank is mislabeled 【44】, duplicating the earlier blank. Question 44's answers are verb endings fitting the first 【44】; question 45's answers are connectives fitting this slot.
  • evidence: jlpt-freq witness (n1-2019.7真题) line 340 numbers this blank 45: 「いつも光って見える。45 裏側に」; answer set for 45 (connectives) fits this slot. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問11 q62 passageA: 「決めるために必要な情をたくさん」 → 「決めるために必要な情報をたくさん」
  • Missing 報; 必要な情 is not a word here, the original reads 必要な情報 (the same passage later says たくさんの情報を合理的に利用して).
  • evidence: jlpt-freq witness (n1-2019.7真题) line 815: 「決めるために必要な情報をたくさん持っていたり」 [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問11 q62 passageB: 「状況に陥陥りやすい」 → 「状況に陥りやすい」
  • Doubled 陥 is a retype duplication artifact; the word is 陥り.
  • evidence: jlpt-freq witness (n1-2019.7真题) line 827: 「という状況に陥り」 with a single 陥 [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問11 q63 passageB: 「しかし、早が必要だとはいっても」 → 「しかし、早さが必要だとはいっても」
  • Missing さ; 早 alone is not the intended noun, the original reads 早さが必要.
  • evidence: jlpt-freq witness (n1-2019.7真题) line 832: 「しかし、早さが必要だとはいっても」 [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問3 q14 question: 「<u>コンパクト</u>な」 → 「<u>コンパクトな</u>」
  • 問題3 underline boundary: synonym options are adnominal, so the trailing な belongs inside the underlined span (same pattern as N1_2020-12 q16 架空の). Without it the answer leaves a dangling な.
  • evidence: structural: 問題3 options replace the underlined adnominal span.
• 問7 q41 passage: 「自分が他人に見せている「自分は、相手に応じて見せる「光り輝いている部分」であって」 → 「自分が他人に見せている「自分」は、相手に応じて見せる「光り輝いている部分」であって」
  • The retype has an unclosed opening 「 before 自分は; grammatically the quoted noun is 「自分」 (the self one shows to others), which is then the topic marked by は and described as 「光り輝いている部分」. Inserting the missing closing 」 after 自分 makes the sentence well-formed (matched quote pair) and restores the parallel 「自分」…「光り輝いている部分」 structure. Flagged as a source typo in the quirks notes (lines 10/13).
  • evidence: Category dropped-char, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問8 q46 passage: 「想像力を鍛えるひとつのポイントになる だろう」 → 「想像力を鍛えるひとつのポイントになるだろう」
  • Stray space inside the verb-auxiliary sequence なる＋だろう; there is no word boundary here, so the space is a retype/layout artifact and the run should read なるだろう.
  • evidence: Category typo, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問11 q62 passageB: 「多角的に物事をとらえられようになることが必要だ」 → 「多角的に物事をとらえられるようになることが必要だ」
  • The pattern is 〜られる＋ようになる ('come to be able to grasp'). The printed とらえられようになる is missing the る of the potential form とらえられる; the grammatical form requires とらえられるようになる.
  • evidence: Category dropped-char, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問11 q62 answers[3]: 「Bはリスクを回避すること ができると述べている」 → 「Bはリスクを回避することができると述べている」
  • Stray space inside the こと＋ができる construction; こと and ができる form one grammatical unit with no word boundary, so the space is a retype/layout artifact.
  • evidence: Category typo, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問13 q69 question: 「再募集に申し 込もうと考えている」 → 「再募集に申し込もうと考えている」
  • Stray space inside the compound verb 申し込む; 申し and 込もう are a single word, so the space is a retype/layout artifact.
  • evidence: Category typo, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問13 q69 question: 「受講料の支払い はどうすればよいか」 → 「受講料の支払いはどうすればよいか」
  • Stray space between the noun 支払い and the topic particle は; no word boundary justifies a space, so it is a retype/layout artifact.
  • evidence: Category typo, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]