解碼Tags：網路大眾分類的語言遊戲與知識意涵

Abstract

本研究以網路大眾分類機制Tagging作為研究對象，有別於以往研究方向，嘗試以分類與知識的角度來看待此新科技，並強調其他研究中缺乏的語境因素。所探討的理論包括傅柯的知識型、知識社會學的社會建構主義等等，試圖解開Tag分類背後的意義。現代知識分類的標準並非天經地義，也不是放諸四海而皆準的，規則的確立受到許多社會歷史因素影響，而Tag分類的出現，開始解構此一嚴密體系，因此產生許多不同以往的分類特性。本研究期望瞭解此新分類科技的思維，以李歐塔的語言遊戲觀念為出發點，藉著觀察Tags在不同類型網站表現出的多樣性，分析Tagging如何運作展現，並探討知識建構和相關語言遊戲呈現的變化和技巧。 研究者首先依據資料型態、資料創作者劃分不同網站類型，選出flickr、YouTube、NICONICO動畫、Blogger和delicious五個網站抽取其人氣檔案作為樣本。並以語境分析的方式進行編碼，發展出人事時地物、主體或背景事物的描述、社群、特殊資訊、專輯、個人、評價、感受、聯想及對話等十大分類類目。將上述十個類別再進一步收攏，可以歸納出四種Tagging主要的分類型態：（1）與主題相關；（2）以公開分享為目的；（3）以私人收藏為目的；（4）情感性標籤。顯示Tag的命名是功能式的，以「功能」而非「事物」做為分類基礎。 研究發現，雖然基本的「人事時地物」標籤仍佔很大比例，顯示傳統分類的影響力深遠，但Tagging方式已不完全按照傳統分類規則，反而常見去階層、主客體界線消融、冗贅等表現。而不同網站的Tagging也呈現不同的特色，不過，雖然文字型網站的確普遍有較多傳統分類，圖像網站較多新型的、情感性標籤，但影響Tagging的是一個更為複雜的網狀關係。除了資料及網站本身的屬性外，研究發現Tagging也受使用者目的與需求影響，但是在不同網站語境中，不會遵循一樣的通則。這顯示Tag命名標準是異質性的，使用者著重的是下標籤時的「語境」和「目的」，而非統一的分類標準。本次研究也發現一些新的Tagging方式，如複雜的詞性、以「情緒」作為搜尋的新節點、藉聯想串連的不同概念、兼具敘事功能及運用各種視覺化、聽覺化的符號等，透過各式各樣的技巧，使用者們得以盡情戲耍能指。 簡言之，在網路空間中，每一個網站就像是一個語言遊戲場，有自己獨特的規則和特性，因此各種Tagging方式，在不同的網站可能有著不同的功能與意義。但研究結果也發現紛亂的Tag會逐漸聚合，顯示使用者對各個網站的規則有自動學習、適應的現象。而有的網站中的使用者甚至目的不在分類，反將Tag欄位當作對話、搞笑的空間，這些看似作怪的Tag並非就不具分類功能，在群眾力量的作用下，它們也可能形成共通的分類。Tag的訂定標準是多重的，拉近了正統知識分類與日常生活經驗之間的距離，將群眾對能指的玩弄，變成知識形塑的一個要件，這種對知識的合法性歷程的扭轉，正是大眾分類最重要的知識意涵。

This study takes the tag classification mechanism as a research object. Different from the previous research, we attempt to treat this new classification from a knowledge perspective, and fill up the lack of situational factors (the language context), observing in a variety of different sites to see the order that tagging present and the meaning of knowledge behind. The theory we concern including episteme by Michel Foucault, the sociology of knowledge by Max Scheler and Karl Mannheim, etc., trying to unlock the meaning behind Tagging classification. Because the standards of modern knowledge is not only natural, nor is it fits-all, the standard established by many social and historical factors, and the emergence of tag classification, began to deconstruct this strict standard, resulting in the classification of many different characteristics compared with the past. We start from Lyotard's concept of language games, trying to observe the diversity of faces Tagging show, and how the classifications and the knowledge construction operate by analysis of different types of sites. According to data types and owners, we first divided tagging system into different four types, then chose flickr, YouTube, NICONICO animation, Blogger and delicious for analysis. We selected popular files and coding them with content analysis and context analysis approach. We developed ten categories, and these ten categories can be concluded into four main types of tagging classification: (1) related to the topic; (2) for the purpose of public sharing; (3) for the purpose of a private collection; (4) emotional labeling. This result shows tag’s naming specifications are based on 'what’s the function' rather than 'what’s the thing'. And after a series of analysis, we found in addition to data type and the properties of the site itself, the user objective and strategy for the use of tagging are also major factors. They cross-impact, molding context as a site’s atmosphere, and will affect the tagging development again. According to the variety of language contexts, the use of tagging will change, too. We also found some brand-new tagging feature. For example, the 'emotion' can be a new node, the different concepts can be linked up by association, the tags can have both classification and narrative features, and can use various visual or auditory symbolic. Users are able to enjoy playing signifiers with such a variety of techniques. In short, in the Internet space each site is like a language game field, and has their own unique rules. Users with sites of this nature use a variety of language strategies to achieve their goals, such as public or private sharing Favorites. We also found that users can learn rules and adapt the phenomena of different sites automatically. In different sites, they may have to take a completely different strategy to achieve the same purpose; some Web site users’ purpose is not classification, but for dialogue or fun. These strange tags are not trouble. By the power of masses, they may also form a common classification. This reversal of the legitimacy process on knowledge is just the most important implications of Tagging.