篆刻是中華文化中流傳已久的一門獨特藝術,於方寸之間納有有章法、篆法與刀法之變化,使得一方小印能各異其趣。本研究主要分為三部分,首先以程序性的方式,從上層結構至下層元件系統化地解構篆刻造形,並針對不同元件定義各種量測尺度並取得數據,客觀清晰的分析篆印風格。第二部分為衍生,依據參數式篆刻形狀文法中的形狀規則與不同風格各異的參數,依序建構篆刻中的各個必須元素,進而重組與模擬特定風格的篆刻。分類為本研究最後一部分,利用第一部分分析結果中的數據,以單因子變異數分析與多變量統計中的判別分析進行統計分析,明白地標出不同風格篆刻之間的界線,並以統計量化角度呼應過往篆刻藝術史料中質性的評論。 研究結果顯示,本研究所發展之篆刻形狀文法,的確可經由衍生的步驟驗證文法的可行性,衍生原風格的篆刻。再者,分類結果也大致符合原有風格分類,且吳昌碩印風中的「留地式」、黃士陵的代表風格「滿白文」與齊白石晚年受「祀三山公」影響而成就的平直方正印文風格,皆可由統計結果知悉。本文中篆刻形狀文法大體架構適合用於描述任一篆刻造形,未來除可解構重組其他風格篆刻,更可擴大運用此一分析方法與邏輯於其他藝術作品。
Chinese seal carving is a unique art which has been around in Chinese history for a long time. The seal style can be affected by the layout of characters, the writing style of strokes and the way of carving. There are three major goals in this dissertation. First, to analyze the form of seals by studying the structure and layout of the seals hierarchically and systematically. The basic measurable elements are extracted according to the historical bibliography. Second, a shape grammar is induced by the analytical result. Different forms of seals can be generated by the grammar by specifying parameters for basic elements in the seal structure. Third, seal forms are converted into measures of basic elements and then classified by statistical methods such as one-way analysis of variance (ANOVA) and discriminant analysis. The classified result is compared to the bibliography to find out how the style is formed for famous seal carving masters. The result of our research indicates that it is possible to generate seal forms by procedural method with parametric shape grammar. The classification for seal styles is quite accurate and acceptable. Especially the well-known styles of seal carving masters Wu Chang-Shuo, Huang Shr-Lin and Qi Bai-Shi are objectively concluded from statistical data. Moreover, the shape grammar can be used to generate seal forms of other style if suitable parameters are discovered.