摘要
人類透過語言傳遞知識。維根斯坦指出，人類透過語言來描繪並認識這個世界。亞理士多德則認為，當我們用語言表達某個判斷時，我們不只表達了一連串的字詞符號，更表達了釵h能呈現世界是怎麼樣的事物(items)。然而，我們的日常語言並不完美。它總是不乏模糊(vagueness)與歧義(ambiguity)，以致於當我們要判斷一個表達是否如實的呈現關於世界的某個事件時，我們所擁有的不完美的語言似乎又變成我們理解這個世界的障礙。此外，從謂述理論(theory of predication)的角度來看，當我們試圖藉助於排中律或二價律來解釋自然語言的謂述(predicating)過程時，這些模糊的表達(vague expression)反而卻造成古典邏輯的困挫。而在這之中，由含糊性所產生最著名的困難，便是堆垛悖論(sorites paradox)。 堆垛悖論是指，當一粒沙子並不足以成堆(a heap)時，則在此不足以成堆的沙上增加一粒沙並不會因此成推；同理，當我們在此無法成堆的沙上再加一粒沙，亦不會使它因此成堆……如此反覆推論下去則會得出一荒謬的結果：不論有多少的沙子均不足以成堆。我們可用堆垛悖論的數學歸納版本來表示—假設
並列摘要
The aim of this thesis is to defend epistemicist theory of vagueness. In the first chapter, I shall analyze the genesis of vagueness and the problems that arise from borderline cases. I will also briefly introduce various approaches toward vagueness; these approaches include Williamson’s epistemicism, Hajek’s degree theory/fuzzy logic, Keefe’s supervaluationalism, and Unger’s nihilism. Next I will thoroughly examine Keefe’s (2000) supervaluationalism and argue that her attempt to apply van Fraassen’s supervaluations on empty name to vague expression is illegitimate. Moreover, her attempt to eliminate semantic vagueness simply misses the point. In the third chapter, I will focus on defending the epistemic view. I will also consider Schiffer’s (1999) and Keefe’s (2000) criticisms of Williamson’s strategy of explaining our ignorance and Mott’s (1998) accusation that the margin for error principle makes iteration of knowledge impossible. In addition, I will try to dispel challenges recently put forth by Dorr (2003), who suggests substituting unassertability with ignorance, and Ray (2004), who attempts to show that epistemicism is self-defeating through his three-way undermining. Last but not least, I will conclude that the epistemic view is the most promising theory of vagueness so far.