M◆fabrik
Look up & Parse · 查詞解析

滿語解析Query

滿語解析能做什麼?→What can Language Query do? →

滿文 詞形 → 詞義 解析器

Note: the dictionary glosses shown are the original Chinese/Japanese definitions from historical Manchu dictionaries — this tool has not built a separate English gloss set, so results below stay in Chinese/Japanese by design, not by mistake.
翻譯模型 v0.6 · 實驗性,不建議用於可靠翻譯translation model v0.6 · experimental, not for reliable use 載入示例
圖例:綠=命中 金=成語/固定搭配(整體命中,後跟成員逐詞虛線卡) 紅=未收錄(人名/擬聲/真缺) 灰藍=存疑分析(不計入命中) · 點詞塊底部彈出詳細詞典釋義
輸入穆麟德轉寫(ū/š 或 v/x 均可)/可直接輸入滿文字,大小寫不限· 成語/固定搭配整體優先 + 逐詞形態切分 + 詞性標註 + 原形還原 + 釋義 · 依季永海《滿語語法》規則 · 詞典來源:《滿漢大辭典》《新滿漢大詞典》《満漢和辭典》三部對接日本東北大學東北アジア研究センター的滿語辭典網站(hkuri.cneas.tohoku.ac.jp);語料詞表對接 manc.hu;人名對接香港金由美教授(Loretta Kim,LEK Research Collective)的 NCNHNL 資料庫;各家資料著作權歸原作者與機構所有,本站僅作檢索之用
Changelog · 更新日誌
2026-09-20詞典來源聲明 — 三部辭典對接日本東北大學滿語辭典網站、語料詞表對接 manc.hu、人名對接金由美教授 NCNHNL 資料庫;著作權歸原作者與機構所有Dictionary source statement — the three dictionaries are accessed through the Manchu dictionary site of Tohoku University (Japan), the corpus vocabulary through manc.hu, the names through Prof. Loretta Kim's NCNHNL database; copyright remains with the original authors and institutions
2026-08-05OCR 頁改版 — 轉寫下方改為完整逐詞解析並顯示詞典命中率;轉寫可就地改錯字,改完即以新文字重新解析;另修譯文配色(淺色主題下不再隱形)OCR page rebuilt — transcriptions now carry a full word-by-word analysis with dictionary hit rate, and are editable in place so the analysis re-runs on your corrections; translation colour also fixed (no longer invisible in light theme)
2026-08-04全站計數口徑統一 — 單位一律用「頁」;詞典原形數改報 65,196(舊值 66,509 把同詞的轉寫變體與無釋義存根各算一條)Counts brought into line — “page” as the sole unit; dictionary base forms now reported as 65,196 (the old 66,509 counted transcription variants and gloss-less stubs separately)
2026-08-01數據庫示例新增《平定準噶爾方略》刻本 × 寫本雙傳本互校(第41–42冊 505 頁):561 處分歧並列標注、不自動合併,查看示例Sample database: Pacification of the Zunghars, printed and manuscript witnesses collated (fasc. 41–42, 505 pages) — 561 divergences juxtaposed and labelled, never merged automatically. See samples
2026-08-01語料庫問答兩修 — 英文提問改用英文作答;《滿文老檔》太宗朝 2,600 頁補上掃描原圖(兩站各 3,803 頁齊全)Corpus Q&A — questions asked in English are now answered in English; 2,600 Taizong-reign pages of the Manchu Veritable Records now have their scans (all 3,803 complete on both servers)
2026-07-29譯文模型 v0.6 — 補回分體裁語域範例;96 句盲評語義充分性 3.34→3.52、意思正確率 79.2%→86.5%(未達統計顯著,仍標示實驗性)Translation model v0.6 — genre-specific register examples restored; over 96 blind-scored sentences adequacy rose 3.34→3.52 and correct-sense rate 79.2%→86.5% (not statistically significant; still marked experimental)
2026-07-26數據庫示例新增《八旗通志初集 · 卷1-11》印刷滿文 4 段共 390 頁(第94–204/207–307/309–410/418–493 頁):三模型共識 OCR + 逐詞解析 + 中英翻譯,查看示例Sample database: added Baqi Tongzhi (initial coll., vol. 1–11), 390 printed-Manchu pages in 4 sections — 3-expert consensus OCR + word-by-word parse + zh/en translation. See samples
2026-07-25滿文閱讀課程《Reading for Manchu》上線 v1.0 — 不做語音、不從字母起步,直接以轉寫掌握語法核心(需口令)Reading for Manchu course v1.0 — no speech, no script drills: straight to the grammar through transcription (password required)
2026-07-25解析頁與 OCR 頁新增「朗讀」功能 v0.1(實驗)Read-aloud v0.1 added to Query & OCR pages (experimental)
2026-07-21M-fabrik 統一站上線 v1.0 — 整合滿語解析、OCR、數據庫示例與使用說明M-fabrik unified site v1.0 — Query, OCR, Database samples and Guide
2026-07滿漢詞典擴充 v1.1 — 併入滿漢人名辭典 11,847 條;解析覆蓋率提升Dictionary v1.1 — 11,847 Manchu-Chinese proper names merged; higher parse coverage
完整更新日誌 →Full changelog →