N1 2021-07 — 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: Cleanly retypeset digital recall version (Chinese prep school 木子小花日本语教室), not a booklet scan: text is crisp and fully legible, with furigana reproduced in the listening section but mostly absent from the written half. Docked for content-fidelity hazards introduced by retypesetting: every page header mislabels the test as 2021年7月N2真题, the 問題13 instruction text describes a farm-park volunteer notice that does not match the actual orchestra-ticket stimulus, a stray '89' is embedded in the Q68 stem, 随康保険証 typo for 健康保険証 in listening 問題1 2番, and the listening cover contradicts itself on page count (Japanese 8ページ vs English 9 pages). Large pale watermark and WeChat ad footer on every page do not impair legibility.

Every section was transcribed twice independently (a full-page pass and a double-resolution half-page pass) and reconciled. Of 13 sections, 1 agreed on all content between the two passes (reconciled automatically, differing only in transcriber notes) and 12 had at least one content difference resolved by a third adjudication pass that re-read the page images. 67 leaf-level differences were examined in total.

Residual OCR risks (30):

• 問題1: Watermark glyphs (木 子 小 花) overlay the answer rows faintly; the small-vs-full kana distinctions (Q6 ちゆ vs じゅ) were legible but rely on a clean digital retype, so a typo in the source itself cannot be ruled out.
• 問題1: This is a 木子小花 recall/retype, so any source-level retype typo is transcribed as printed and not corrected.
• 問題2: All four （　　） blanks (q7-12 stems) print as a single wide blank run; normalized to （　　） per convention regardless of printed width, which is correct but masks the original blank widths.
• 問題2: しょんほり is preserved as a printed typo; a downstream answer-key/normalization pass should be aware the standard form is しょんぼり.
• 問題3: Q14 doubled period 「。。」 is preserved as printed per convention; it is almost certainly a retype typo in the source, not a meaningful glyph.
• 問題3: Source question numbers are printed boxed (14-19 in boxes) despite the task note expecting plain numbers; number boxing is not part of any transcribed text field, so no content impact.
• 問題4: All six questions are word-usage items with no furigana, blanks, or tables, so structural risk is minimal; only the Q24 trailing な and the Q22 doubled period are subtle and both were verified directly against high-res crops.
• 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: The 問題6 instruction-line slot blank ( ★ ) is rendered with full-width low lines ＿ flanking the ★; the convention's explicit ＿＿＿ rule covers the '＿＿の言葉' instruction blank rather than the star-slot instruction, so the ＿★＿ rendering is a reasonable faithful choice but is a judgment call on glyph count/width.
• 問題6: Watermark overlays (小花 / 日本语教室) cross several lines but did not obscure any transcribed text in mondai6.
• 問題7: Light watermark overlaps parts of the 【41】 and 【43】 lines; readings were confirmed at 8x zoom but a faint stroke could be obscured.
• 問題7: The 【43】-paragraph ending glyph cluster 'なんかは….' is rendered small; it is one 3-dot ellipsis plus a half-width period at best reading, but the exact dot count of the leader is slightly fuzzy.
• 問題7: Question count and structure (4 questions, single passage) match the expected 問題7 = 41-44 spec; no structural ambiguity remains.
• 問題8: The source normalizes some punctuation oddly: passage (4) is printed with Chinese-style commas '，' and a middle dot in '専門家・研究者' / '印刷・販売'; both transcribers rendered commas as Japanese '、' and kept '・'. Followed that agreed normalization; the raw print is half Chinese-comma.
• 問題8: Email <p>/<br> structure for passage (2) is a judgment call. Hard line breaks vs prose-wrap were distinguished by whether a line reaches the right margin; a couple of breaks (e.g. before 'すでに') are short author breaks rendered as <br>. A reviewer applying a stricter 'blank line = <p>, everything else collapses' rule could differ on the <br> placement.
• 問題8: '環境問題' renders with a 間-like 問 glyph in the Q48 stem at this font; read as 問題 (standard), consistent with both transcribers.
• 問題9: The double-prime quotation glyph around わかる in passage (2) para 4 ('" わかる"') is rendered with an ASCII double-quote and surrounding spaces; the printed mark is a ditto/double-prime style glyph and exact character + spacing is a judgment call.
• 問題9: The single em-dash in passage (2) para 4 ('「わかる」―どうでもいい') is transcribed as one U+2015; both transcribers used a full-width hyphen-minus (U+FF0D). Glyph identity is slightly ambiguous at this resolution.
• 問題9: Passages (2) and (3) required heavy reconstruction because both OCR passes diverged from the print in many places; while every clause was confirmed on the -t/-b crops, a residual risk of a missed particle remains given the density of retypeset errors.
• 問題10: The 「というように…。」 trailing mark: the print shows one 3-dot ellipsis group then a single small low dot. Rendered as …。 (ellipsis + full-width period); the final glyph could conceivably be a half-width period in the Chinese retype, but a full-width 。 is the most reasonable reading.
• 問題10: The dash in 「こう考える―会社に対する」 is a single short dash in the source (not a 2-em dash); transcribed as a single ― (U+2015), matching both transcribers, but the exact intended dash glyph is uncertain.
• 問題10: Question numbers 58-61 are printed inside small boxes here (not plain as the section note suggested); this does not affect transcription but is noted in case other sittings differ.
• 問題11: The アビール/アピール spelling is a deliberate retype typo in the source; preserved as printed per conventions. A consumer normalizing katakana would want to know アピール is intended.
• 問題11: Source is a Chinese prep-school re-typeset Word→PDF whose header mislabels the level as N2; content (boxed 62/63, two-passage 問題11 format) confirms it is N1 2021-07.
• 問題12: The passage is a long single block typeset by a Chinese prep school; while every disputed span was re-read at double resolution, a single-character misread elsewhere in the dense text cannot be 100% excluded.
• 問題12: 「生きっづける」 is an apparent source typo (intended 生きつづける); transcribed as printed per conventions but flagged in notes.
• 問題12: The source itself is internally inconsistent (著者 vs 筆者 across stems; passage's ②「子どもらしい」 vs the narrative line 「子どもらしくない」); these are genuine source artifacts, not OCR errors, and are preserved.
• 問題13: Table-header parens were judged half-width by comparing their tight glyph to the clearly wider full-width （A席） in Q69; at this raster resolution a deliberate half-width choice is plausible but not 100% certain.
• 問題13: Address-line spacing (full-width spaces before チケットオフィス and 〒, half-width before 東月市) is inferred from gap widths in p24-b; whitespace width is inherently the least certain glyph class.
• 問題13: The （　　）blank in '各席種の（　　）の中の金額' was rendered with the fill-in-blank convention (full-width parens + two full-width spaces); the printed blank's exact internal width is approximate, but the parens-blank (not 表中) reading is certain.

Source-text quirks the OCR transcribed verbatim (printed typos/recall artifacts; the ones judged to be retype errors were corrected in the layer below) — 23 noted at OCR time:

• 問題1: Q4: the printed underline extends under the sentence-final 。 as well (憤った。), so the period is included inside <u>…</u> as printed; confirmed in both the full-page render and the high-resolution half-page.
• 問題1: Q6 options 1 and 2 use full-size kana (ちゆ／ちよ) while options 3 and 4 use small kana (じゅ／じょ); verified at high zoom.
• 問題2: Q12 option 1 is printed as 「しょんほり」 (unvoiced ほ, no dakuten); confirmed against the double-resolution p03-t crop. This is a retype typo for the standard 「しょんぼり」; transcribed exactly as printed.
• 問題3: Q14: the stem ends with a doubled full-width period 「だったそうだ。。」 — transcribed as printed (likely a retype typo).
• 問題4: Q20 answer 1 begins 'これば' (likely a retype typo for 'これは'); transcribed as printed.
• 問題4: Q22 answer 3 ends with a doubled full-width period '。。' (彼に仕事を頼んだら、露骨に嫌な顔をされた。。); transcribed as printed, likely a retype typo.
• 問題5: Q29: source prints 通動 (likely a retype typo for 通勤); transcribed as printed.
• 問題5: Q30: stem ends without a period ("…という意味です" with no 。) as printed.
• 問題5: Q35: the sender name reads 中為; transcribed as printed (uncertain reading, possibly a retype artifact).
• 問題5: Q29: source prints「通動してくる」(動 instead of 勤) — likely a retype typo for 通勤; transcribed as printed.
• 問題5: Q30: stem ends without a closing 。 in the source (「…意味です」 with no period); transcribed as printed.
• 問題5: Q35: sender name printed as「中為」which is an unusual surname (為 may be a retype slip); transcribed as printed.
• 問題6: Slot star positions after order-audit cleanup: q36-q40 have ★ in the third slot ( _ _ ★ _ ).
• 問題7: 「云う」「云って」「云いながら」「云えば」 are used throughout (non-standard kanji for いう); transcribed as printed.
• 問題7: Inconsistent box-shape wording in the source: '長細い箱' appears three times (paragraphs 1, 5, 6) while '細長い箱' is used in paragraph 3 — transcribed as printed.
• 問題8: Passage (1) ends '織り込みずみだ' (printed 'ずみ' rather than the more standard 'すみ'); transcribed as printed.
• 問題8: Passage (2) is a bookstore email shown below a lead-in line and inside a boxed frame; the lead-in '下は、書店から送られてきたメールの内容である。' is the first <p>, blank-line-separated blocks are <p>s, and the author's hard line breaks within a block are <br>. Email address printed as tyuumon@bukkunetto-a-syoten.jp; transcribed as printed.
• 問題9: Underlining of reference phrases is inconsistent in this retypeset source and was transcribed per-location: passage (1) does NOT underline ①かんたんである but DOES underline the ② and ③ spans; passage (2)/(3) DO underline the ①/② passage spans. Question stems 49-51 print the underline, but stems 52, 55, 56 print only the circled mark (52/55) or plain phrase (56) with NO underline. Both transcribers over-applied <u> to stems 52, 55, 56; corrected here against the images.
• 問題9: Passage (2): both transcribers garbled the third paragraph. The image reads 'つまりそれは、「わかる」のである。「わかる」は、「わからない」を解明するためのヒントである。つまりは、（　②　）ということである。' — transcriber A's spurious '翅' came from misreading this run; the （　②　） blank has no surrounding 「」.
• 問題9: Likely retypeset typos transcribed as printed: passage (1) q49 option 1 '雌の翅の色は光であり直進するから'; passage (2) 'こんなくだらないものの答が全体像の解明につながるはずはない」' has an unbalanced quote run (one 「 vs one 」).
• 問題11: Passage B prints "アビール" twice (in the first paragraph 「最大限にアビール」 and the last sentence 「ということをアビール」) where standard Japanese would be "アピール"; verified against the double-resolution image (clear dakuten ビ, not handakuten ピ) and transcribed as printed (likely a retype typo).
• 問題12: Retype typo in the passage (transcribed as printed): 「生きっづける」 — the printed text shows a small っ at line end followed by づける on the next line, yielding 生きっづける; clearly intended 生きつづける. (Transcriber A read this as 生きづづける; the glyph is a small っ, not づ.)
• 問題12: Inconsistency in the source (transcribed as printed): question 66's stem quotes the underlined phrase as ②「子どもらしい」 (matching the passage's ②「子どもらしい」), while questions 64/65/67 stems use 筆者 but question 66 uses 著者. Both 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 q14 question: 「<u>寡黙</u>な人」 → 「<u>寡黙な</u>人」
  • The choices are complete adnominal clauses, including the correct 口数が少ない. Leaving な outside the tested span produces 口数が少ないな人; the full target is 寡黙な.
  • evidence: structural: 問題3 choices replace the underlined span.
• 問3 q17 question: 「<u>ろく</u>に説明」 → 「<u>ろくに</u>説明」
  • The choices are complete adverbials, including the correct 大して. Leaving に outside the tested span produces 大してに説明; the full target is ろくに.
  • evidence: structural: 問題3 choices replace the underlined span.
• 問1 q4 question: 「<u>憤った。</u>」 → 「<u>憤った</u>。」
  • Print underlines only 憤った; the closing 。 is outside the underline
  • evidence: underline-boundary-vs-image
• 問3 q16 question: 「<u>ずれ込み</u>そうだ」 → 「<u>ずれ込みそうだ</u>」
  • Print underlines ずれ込みそうだ (matching options 〜なりそうだ); our span dropped そうだ
  • evidence: underline-boundary-vs-image
• 問2 q12 answers[0]: 「しょんほり」 → 「しょんぼり」
  • しょんほり is a non-word; the intended mimetic is しょんぼり. 問題2 options are all real words (ぎくしゃく, ぼんやり, うろちょろ are genuine mimetics), so the unvoiced ほ is a retype typo, not a deliberate distractor.
  • evidence: Confirmed exact substring at answers[0] of number 12. Other three options are real mimetics; this is a vocabulary fill-in, not a usage-distractor, so all options must be real words. しょんぼり is standard. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問5 q29 question: 「自転車で通動してくる」 → 「自転車で通勤してくる」
  • 通動 is not a word; intended word is 通勤 (commute). It sits in the carrier sentence (not a tested answer), so 動 for 勤 is a retype typo.
  • evidence: Confirmed exact substring in the number 29 question. 通勤 is standard; 通動 is a non-word. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問4 q20 answers[0]: 「これば相手のチームを倒す」 → 「これは相手のチームを倒す」
  • これば is a non-word kana artifact; the sentence-initial topic marker これは is intended. 問題4 distractors misuse the tested word (絶滅), not function words, so this is not deliberate content. Only ば→は changes.
  • evidence: Confirmed exact substring at answers[0] of number 20. これば has no valid reading; topic marker これは is unambiguously correct. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問3 q14 question: 「な人だったそうだ。。」 → 「な人だったそうだ。」
  • Doubled full stop 。。 is a retype artifact; the official exam prints a single 。.
  • evidence: Confirmed exact substring at end of number 14 question. Official JLPT never prints double periods. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問4 q22 answers[2]: 「に嫌な顔をされた。。」 → 「に嫌な顔をされた。」
  • Doubled full stop 。。 is a retype artifact; the official exam prints a single 。.
  • evidence: Confirmed exact substring at answers[2] of number 22. Official JLPT never prints double periods. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問12 q64 passage: 「私たちは生きっづけるために」 → 「私たちは生きつづけるために」
  • 生きっづける contains a stray small っ; intended form is 生きつづける (a line-break/typesetting retype artifact). Only the small っ→つ changes.
  • evidence: Confirmed exact substring in the mondai12 passage. 生きっづける is ungrammatical; 生きつづける is the intended verb. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問11 q62 passageB: 「最大限にアビール」 → 「最大限にアピール」
  • アビール is a non-word; intended katakana loanword is アピール (appeal). ビ for ピ is a dakuten/handakuten retype typo.
  • evidence: Confirmed exact substring in passageB. Standard loanword is アピール. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問11 q62 passageB: 「ということをアビール」 → 「ということをアピール」
  • アビール is a non-word; intended katakana loanword is アピール (appeal). ビ for ピ is a dakuten/handakuten retype typo.
  • evidence: Confirmed exact substring in passageB (second occurrence). Standard loanword is アピール. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問10 q0 passage: 「（注1）に考」 → 「に（注1）考」
  • （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10 q0 passage: 「（注2）の価」 → 「の（注2）価」
  • （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12 q64 passage: 「持っていない子をのけ（注1）ものにしたり」 → 「持っていない子をのけもの（注1）にしたり」
  • （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6 q39 question/correctOrder: 「全国高校バスケットボール大会で、惜しくも _ ★ _ _ しばらくの間、ぼう然としていた。; correctOrder null」 → 「全国高校バスケットボール大会で、惜しくも _ _ ★ _ しばらくの間、ぼう然としていた。; correctOrder [3,0,2,1]」
  • The accepted assembly is あと一歩というところで優勝を逃した選手たちは控え室に戻っても..., order 4132. Since the answer key gives q39 answer 3, ★ belongs in the third slot.
  • evidence: input/N1_2021-07/5Mcv4He-jlpt-freq-exam-and-answers.txt lines 2031-2033: q39 答案 4132 and assembled sentence. [order-conflict cleanup, 2026-06-24]
• 問6 q40 question/correctOrder: 「私が接客するにあたって _ ★ _ _ 自分は何をすべきかということだ。; correctOrder null」 → 「私が接客するにあたって _ _ ★ _ 自分は何をすべきかということだ。; correctOrder [0,3,1,2]」
  • The accepted assembly is 常に考えているのは、お客様が求めていることは何であって、それに応えるために..., order 1423. Since the answer key gives q40 answer 2, ★ belongs in the third slot.
  • evidence: input/N1_2021-07/5Mcv4He-jlpt-freq-exam-and-answers.txt lines 2048-2050: q40 答案 1423 and assembled sentence. [order-conflict cleanup, 2026-06-24]