近年來以臺灣為主體的社會、歷史、教育、文化研究蓬勃發展，人們日益重視日治時代的臺灣文獻資料。圖書文獻必須經過有系統的組織整理、分門別類之後，方能提供研究者查檢使用的便利性。 本研究旨在探討日治時代臺灣文獻的分類體系，採用個案研究法，並透過深度訪談，以了解國立中央圖書館臺灣分館、國立臺灣大學圖書館、國史館臺灣文獻館的臺灣文獻資料的內容形式、收藏情況、所採用的分類體系及應用情況，是否能提供現代分類系統編訂時的啟發。 根據研究發現，提出結論如下：1. 日治時代臺灣文獻特徵；2. 臺灣文獻資料的館藏管理；3. 日治時代臺灣文獻資料分類體系；4. 臺灣文獻資料分類應用。 本研究建議：1. 舊藏資料的分類體系有其歷史意義和價值，值得保留沿用；2. 中國圖書分類法臺灣相關類號之修訂建議；3. 建立不同分類系統的對照表，便於跨分類系統和提供主題檢索的機制；4. 透過數位典藏的方式，將不同典藏單位的臺灣文獻資料實現網路上的聚合，強化分類系統的應用範疇；5. 舊藏資料的分類體系可以加以推廣，讓使用者容易掌握館藏資料之主題內容和特色。 / Taiwan Study is a field that has been growing rapidly in the past score, especially when it comes to on social, historical, educational, and cultural issues of the Japanese Colonial Period (1895~1945 C.E.). Since then, library materials and archives created during that time have become more and more important for researchers of today. The functions of classification include shelving, browsing, and searching. They help users to access to those materials. The purpose of this research is to discover the value of classification system on materials of the Colonial Period. The research method of the study includes “Case Study” and “In-depth Interview”`. This study dwells on the content, format, classification systems, and applications of Taiwan materials in the three cases as follows: National Taiwan Library, National Taiwan University Library, and Taiwan Historica. Below are the findings: 1. Characters of Taiwan materials of the Japanese Colonial Period. 2. Collection management of Taiwan materials. 3. Classification systems of Taiwan materials of the Japanese Colonial Period. 4. Application of classification systems. The study makes final suggestions as follows: 1. Reserve the practical and historical values of classification systems. 2. Take into consideration the “New Classification Scheme for Chinese Libraries” as presented in the study. 3. Establish conversion tables between various classification systems and a mechanism for subject search through catalogues. 4. Aggregate Taiwan materials by creating digital archive and extend the scope of application in classification systems. 5. Promote the old classification systems to allow users to know more about content and features of Taiwan materials.

檔案控制層次分為管理、全宗、系列、案卷、件等五個層次，是由美國國家檔案暨文件署之學者Oliver Wendell Holmes在1964年提出「檔案編排-五種不同的控制層次」一文的說法。 從2001年起，國內各典藏歷史檔案機構已就其重要的歷史檔案，引進採用最新發展之「檔案描述編碼格式」(Encoded Archival Description，簡稱EAD)及「國際檔案描述標準」(General International Standard Archival Description，簡稱ISAD(G))等標準，進行數位典藏計畫，使檔案控制層次在歷史檔案編排之應用，成為數位典藏計畫執行期間及本研究階段之重要研究課題。 本研究採用「個案研究法」與「深度訪談法」進行研究，對5所典藏歷史檔案機構進行個案研究，包括：國立故宮博物院、國史館、國史館台灣文獻館、中央研究院歷史語言研究所及近代史研究所，以了解檔案控制層次之應用成果；另深度訪談法則是對機構內檔案編排人員進行訪談，以搜集參與數位典藏計畫前後，檔案控制層次在歷史檔案編排之應用情形資訊，及檔案描述有關之控制層次內容。 根據本研究，有下列結論：一、各機構均會參酌最新理論與應用先例進行檔案編排；二、對於檔案控制層次除運用在實體歷史檔案編排外，並也已逐漸延伸到數位檔案影像編排上；三、後設資料需求規格書中之控制層次名稱更為明確；四、控制層次邏輯可以作為製成檢索工具基礎及應用在數位檔案之檔名命名上；五、層次關係以樹狀結構呈現；六、多數機構著重在較低層次的描述；七、彙整出各機構檔案編排之層次圖表。 最後對前述研究結論提出建議：一、未將檔案控制層次運用在歷史檔案編排者，應加以調整；二、控制層次之名稱與內涵應一致；三、運用檔案控制層次方法後，應加以彙集出版或作為網路上瀏覽檔案的輔助索引；四、建立描述計畫時，應著重在較高層次的描述；五、各機構應加強訂定檔案描述規則與多層次描述規範。 / The levels of control in archival arrangement can be classified by depositories, fonds, series, files and items ,etc. This concept had been proposed by Oliver Wendell Holmes in 1964, in his article: “Archival Arrangement -- Five Different Operations at Five Different Levels”. Since 2001, many public archive institutions which are major holders of historical materials in Taiwan, early or late participated in the “National Digital Archives Program” (NDAP). They have adopted “Encoded Archival Description” (EAD) and “General International Standard Archival Description” ( ISAD(G)) as the standards of archive digitization. Therefore, the levels of control applied in the arrangement of historical archives have become the main topics of NDAP and relative studies. This study has adopted the “case study method” and the “in-depth interview method.” We make case studies of five major historical archive institutions in Taiwan -- the National Palace Museum, the Historica Academia, the Taiwan Historica, the Institute of History and Philology and the Institute of Modern History, especially regarding the levels of control applied in the arrangement of historical archives before and after these institutions participated in NDAP. On the other hand, we interviewed the archivists of the above institutions, in order to understand their work of archival description based on the levels of control. We reached the following conclusions: 1. The above-mentioned institutions adopt the newly theory and practice in archival arrangement every now and again. 2. The levels of control not only have been applied in the arrangement of physical archives, but extended to digital archives. 3. The levels of control have become more definite in the specification of metadata requirement. 4. The logic of the levels of control can be utilized as the basis of finding aids and : the naming rule of digital files. 5. The hierarchical model of the levels of control has been presented in a kind of tree structure. 6. Most institutions focus their efforts on lower level descriptions. 7. For further research, we establish a summary table of the levels of control adopted by the above-mentioned institutions. Finally, we do have some suggestions: 1. For every historical archive holder, it is a wise choice to adopt the levels of control to arrange the historical materials 2. The naming on the levels of control and its contents must be consistent. 3. The levels of control applied in the arrangement of historical materials should be available on-site to user as auxiliary index. 4. We’d better concentrate more efforts on higher level description while planning for archival description. 5. Each public archive institution should make every effort to enhance the establishment of Chinese archival description rules and multiple description specifications.