N1 2018-12 — 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 4/5: Born-digital render of the school's retypeset document: crisp, high-contrast horizontal text, fully legible on every page. Every page carries light-gray 帝京日語 logo watermarks plus repeated contact-info watermark text (phone/QQ/wechat/URL) behind the body text, and promotional headers/footers; these never impair legibility. Main quality risk is content fidelity, not legibility: the publisher explicitly labels it a roughly-proofread rush version with many errors, and it is not the original JLPT typesetting (no furigana, no boxed numbers).

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. 42 leaf-level differences were examined in total.

Residual OCR risks (23):

• 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: 「子どもはい先生」(passage 3) and 「鮮度が重視の」(passage 1) are almost certainly retype typos in the source Word document, but are transcribed exactly as printed per convention; correctAnswerIndex left null.
• 問題8: Email recipient/signature line grouping (<br> within one <p> vs separate <p> per line) is a formatting judgment not explicitly pinned by the conventions; chose <br>-joined blocks because the lines are consecutive with no paragraph break, while the枠外 lead-in line is a separate <p>.
• 問題8: Name 戸田 vs 戸竹: the glyph is a clean digital 田 (boxed) in the email box; high confidence it is 戸田 despite A reading it as 竹 to match the company name 沖竹.
• 問題9: The source is an admitted rush retypeset (抢先版本) with acknowledged errors; passage wording is transcribed verbatim and may diverge from the official exam text (e.g. 共体験 likely should be 追体験), but matches what is printed.
• 問題9: Passage citation lines are not present in this section as printed (no （…による） source line); none was added.
• 問題9: The instruction points marker (2*9) is intentionally omitted per the no-points-marker convention.
• 問題10: （注2）'作用のなすがままに：' colon and （注1）'生息生存' no-colon: confirmed visually, but the source is an admittedly rushed Chinese prep-school retype, so the no-colon （注1） is itself likely an upstream typo (preserved per convention).
• 問題10: Instruction-line annotation （3*4） is a scoring/answer-key artifact and was intentionally dropped per conventions; if a later pass needs it, it lives only in the notes.
• 問題11: Source is a self-described rush/proof version (抢先版本, 仍有很多纰漏) and may itself contain retype typos relative to the official exam; per conventions these are transcribed as printed and not corrected.
• 問題11: Passage A paragraph 1 ends with 事実。 (a deliberately abrupt nominal ending in the source). This is faithfully transcribed and read as such, but is a plausible source typo for 事実です。
• 問題12: Paragraph 3 "あふれてたため" is a printed retype typo; transcribed verbatim. Confirmed the line truly ends in "あふれてたた" + next line "め" rather than a smudge, but the doubled た is unusual.
• 問題12: Instruction "(34)" exclusion is a judgement call: it IS printed on the instruction line (not in a separate table). Chose to exclude as school scoring noise; if the project later prefers verbatim instruction lines, switch instruction to include " (34)".
• 問題12: "モン" in paragraph 6 is printed as-is (likely a typo for モノ) and kept verbatim.
• 問題13: 子供（4-7歳）dash: rendered as a straight short dash; represented as ASCII '-'. If the source glyph is a full-width 全角ダッシュ '－' the byte differs, but it is clearly NOT the wave dash 〜 used in 'a.〜c.' (which remains 〜).
• 問題13: Half-width vs full-width space between 施設課 and 自然の宿担当: judged half-width from a visibly smaller gap than the full-width 役所→施設課 gap; low-confidence at this resolution.
• 問題13: Heading modeled as a single <br>-joined <p>; the two centered lines could alternatively be two <p> paragraphs — a presentational judgment, not a text difference.

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: q6 option 1 きょたん: retype/全文 give きょたん; primary OCR had きよたん (typo). Used きょたん.
• 問題5: q33 speaker labels: retype shows 母/父; primary OCR had a garbled '役'. Used 母. option 1 やってるって confirmed.
• 問題6: 問題6 ★ slot position UNCERTAIN: both witnesses are OCR scans that rendered the printed ★ as garbles (太/文/支/3) and did not cleanly preserve the four blank slots. The ★ has been placed in slot 3 of four as a default; the true printed ★ position for each item could not be verified from the text witnesses and should be re-checked against an image scan.
• 問題7: Title 装飾という架け橋: both witnesses garbled the lead character (礎/村/上雲飾); reconstructed as 装飾.
• 問題9: All three passages are rendered in very small, dense type; passage text was verified line-by-line against the double-resolution half-page renders and is high-confidence. Transcriber B's passage bodies were heavily garbled by OCR and were discarded in favor of A's (image-confirmed) versions.
• 問題9: This is the 成都帝京 rush version (抢先版本), which the source itself admits has many errors; some passage sentences read as awkward but were transcribed exactly as printed.
• 問題9: Passage (1): closing bracket mismatch as printed — 「音楽の聴き方』 opens with 「 but closes with 』 (image-confirmed).
• 問題9: Passage (1): 意味/意義 uses an ASCII slash as printed (normalized per convention).
• 問題9: Passage (2): retype space artifacts inside text, e.g. 機密情報の漏洩 など and 既存業務だけで は and 成長産業へ緩 やかに; transcribed these gaps as printed single spaces (image-confirmed).
• 問題9: Passage (1) Q52 option 4 reads 共体験している (likely intended 追体験 in the original source, but printed as 共体験); transcribed as printed.
• 問題9: Passage (1) Q52 option 1 (音楽を最後まで聴いて印象に残るかどうか) has NO trailing 。, unlike options 2–4 which end in 。; transcribed as printed (image-confirmed).
• 問題9: Q56 options 2 and 4, and Q57 options 2 and 4, all end in か rather than から, unlike options 1 and 3 in each. Verified character-by-character against the high-res renders (likely retype typos); transcribed exactly as printed.
• 問題9: Note colons are inconsistent in the source: 注1 in passages (1) and (3) use a half-width colon (刷り込まれる: ここでは / 空虚だ: むなしい) while other 注 use full-width ：; transcribed each as printed.
• 問題9: Passage (3): 空虚（注1)で / 心に平安（注3)を — closing parenthesis around the 注 reference is printed half-width ) while the opening is full-width （; transcribed as printed (image-confirmed).
• 問題9: Passage (2): （注4）主務先ここでは、本業の勤務先 — missing the separator colon that the other 注 glosses have; transcribed as printed.
• 問題9: Passage (2) 注1 gloss is printed モチペーション (with handakuten ° → ペ) even though the body text uses モチベーション (ベ); transcribed each as printed (image-confirmed).
• 問題9: Circled reference marks ①② before underlined spans transcribed as printed text immediately before the <u> span.
• 問題10: （注1）is printed as『生息生存』with no colon/gloss, unlike the other 注 which use the『語：説明』form. Verified against the double-resolution p15-b image. Transcribed exactly as printed; likely a retype error in the source.
• 問題10: The phrase そう簡単ではありません is underlined both in the passage (paragraph 3) and in the Q60 stem; verified against the double-resolution p15-t and p16-t images.
• 問題11: Passage A was small in the full-page render; verified character-by-character against the double-resolution top-half image (p17-t.png cropped/upscaled). Transcriber B's passageA was a substantially incorrect reconstruction (B's own notes admitted low confidence here); the adjudicated passageA matches transcriber A.
• 問題12: Paragraph 1 reads "モノにではなく、そこに込められた「思い」こそが…" (verified against double-resolution image).
• 問題12: Paragraph 3 is printed "…サービスがあふれてたため、…" — the line ends "あふれてたた" and the next line begins "め"; this is a retype typo for "あふれていたため" but transcribed exactly as printed.
• 問題12: Paragraph 6 begins "ただ、モンによって自分を他と差別化する…" — "モン" appears to be a retype typo for "モノ"; transcribed as printed. Reads "…自分を個性化しようとすればするほど本当の自分の色を消し去り、逆に没個性化していくという矛盾である。".

Corrections applied (19)

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

• 問3 q14 question: 「<u>スライス</u>して皿に」 → 「<u>スライスして</u>皿に」
  • All four choices are complete て-form phrases, including the correct 薄く切って. Leaving して outside the tested span makes their substitutions ungrammatical; the full target is スライスして.
  • evidence: structural: 問題3 choices replace the underlined span.
• 問8 q48 passage: 「子どもはい先生」 → 「子どもはいい先生」
  • 「い先生」is a non-word; intended phrase is「いい先生」(children spot a good teacher). Dropped い is a retype artifact.
  • evidence: Witness line 356: 「子どもはいい先生を見抜きます」. File passage reads 子どもはい先生を見抜きます. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q56 answers[1]: 「長い人生で積み上げてきた経験があるか」 → 「長い人生で積み上げてきた経験があるから」
  • 「なぜか」question; sibling options 0 and 2 end in から. Truncated か is a retype artifact.
  • evidence: Witness line 494: 「長い人生で積み上げてきた経験があるから」; kaisetsu line 1267 quotes same. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q56 answers[3]: 「時間が早く経つのを知っているか」 → 「時間が早く経つのを知っているから」
  • 「なぜか」question; parallel options end in から. Truncated か is a retype artifact.
  • evidence: Witness line 497: 「時間が早く経つのを知っているから」. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q57 answers[1]: 「なにひとつ思いどおりにできず悩んでいたか」 → 「なにひとつ思いどおりにできず悩んでいたから」
  • 「なぜか」question; parallel options end in から. Truncated か is a retype artifact.
  • evidence: Witness line 502: 「なにひとつ思いどおりにできず悩んでいたから」. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q57 answers[3]: 「おもしろくないことばかりやれといわれたか」 → 「おもしろくないことばかりやれといわれたから」
  • 「なぜか」question; parallel options end in から. Truncated か is a retype artifact. To preserves file's kana おもしろくない.
  • evidence: Witness line 504: 「面白くないことばかりやれといわれたから」 (ends in から). [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q53 passage: 「主務先ここでは」 → 「主務先：ここでは」
  • The （注4） gloss is missing its separating full-width colon; all sibling glosses in this passage use ：(e.g.（注5）コンサルティング：ここでは…).
  • evidence: Witness line 448: 「（注4）主務先：ここでは、本業の勤務先」; parallel gloss formatting in same passage. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問12 q65 passage: 「あふれてたため」 → 「あふれていたため」
  • 「あふれてた」is missing い; the source reads「あふれていた」.
  • evidence: Witness lines 658-659: 「身の周りに手作りのモノやサービスが…あふれて／いたため」; Google Books たのしい不便/福岡賢正 confirms あふれていた. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問12 q67 passage: 「モンによって」 → 「モノによって」
  • 「モン」is a retype typo for「モノ」; the passage uses モノ 6 times elsewhere and モン only here.
  • evidence: Witness line 674: 「モノによって自分を他と差別化する」; Google Books たのしい不便/福岡賢正 confirms モノ. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問13 q69 passage: 「nakahashi.lg.ip」 → 「nakahashi.lg.jp」
  • Japanese local-government domains use .lg.jp; .ip is not a valid TLD. The i-for-j is a transcription/OCR artifact.
  • evidence: Witness line 782: 「http://www.nakahashi.lg.jp/shizennoyado」. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問8 q46 passage: 「鮮度が重視の生鮮食品」 → 「鮮度重視の生鮮食品」
  • 「重視」is a noun here, so 「鮮度が重視の…」is ungrammatical; the natural and intended compound is 「鮮度重視の生鮮食品」(fresh food where freshness is emphasized). The inserted が is a retype error.
  • evidence: Category added-char, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問13 q70 question: 「10月1日に「中橋自然の宿に泊まりたいと考えている。予約が申し込める期間はいつからいつまでで、区民料金で宿泊する」ために」 → 「10月1日に「中橋自然の宿」に泊まりたいと考えている。予約が申し込める期間はいつからいつまでで、区民料金で宿泊するために」
  • The facility name is quoted as a unit 「中橋自然の宿」 everywhere else (passage and Q69). Here the closing 」 was dropped after 宿 and a spurious 」 inserted after 宿泊する. Restoring 」 after 宿 fixes the corrupted quotation.
  • evidence: Category typo, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問9 q56 passage: 「空虚（注1)で」 → 「空虚（注1）で」
  • Body note marker uses a half-width closing paren 「)」 while every other 注 marker in the exam (including this passage's own gloss line 「（注1）空虚」) uses the full-width 「）」. The mismatch is an OCR/retype artifact.
  • evidence: Category orthography, confidence medium. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問9 q56 passage: 「平安（注3)を」 → 「平安（注3）を」
  • Same half-width closing-paren OCR artifact as 注1 in this passage; the exam uses full-width 「）」 consistently.
  • evidence: Category orthography, confidence medium. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問9 q53 passage: 「人事部·人事白書」 → 「人事部・人事白書」
  • The separator is a Latin MIDDLE DOT (U+00B7) rather than the Japanese full-width katakana middle dot ・ (U+30FB) used in Japanese title separators. OCR/retype artifact.
  • evidence: Category orthography, confidence medium. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問13 q70 passage: 「電話：015-539-15150」 → 「電話：015-539-1515」
  • The last group has five digits (15150) while the companion fax number 015-539-1516 has four; the 015-539-XXXX pattern is broken by an extra digit, a typical retype error (phone/fax usually differ only in the final digit, 1515/1516).
  • evidence: Category typo, confidence medium. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問10 q0 passage: 「（注2）のなすがままに置」 → 「のなすがままに（注2）置」
  • （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 q0 passage: 「（注1）のなさ」」 → 「のなさ（注1）」」
  • （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 q40 question/correctOrder: 「自分が _ _ ★ _ 経験として受け入れられると思う。; correctOrder null」 → 「自分が _ ★ _ _ 経験として受け入れられると思う。; correctOrder [1,3,2,0]」
  • The accepted assembly is 自分が決めたことならそれはそれでひとつの経験..., order 2431. Since the answer key gives q40 answer 4, ★ belongs in the second slot.
  • evidence: input/N1_2018-12/5Mcv4He-jlpt-freq-exam-and-answers.txt lines 1167-1169: q40 正解/答案 4 and assembled order 2 決めたことなら 4 それは 3 それで 1 ひとつの. [order-conflict cleanup, 2026-06-24]