後現代文學評論家羅蘭．巴特〈作者之死〉與密謝．傅柯〈作者為何〉兩篇文章，成功扭轉了傳統上大眾對於「作者」的觀念。兩位學者皆同意，當作者的影響力消失時，「文本」本身則能夠被讀者豐富地、多層面地解讀。巴特與傅柯的觀念源自於俄國文學評論家巴赫金。巴赫金所提出的觀念與諸多後現代文學批評觀點有濃厚的相關性，尤其是「作者之死」，促成了當代「讀者反應理論」以及相關支派的脈絡與興盛。巴赫金提出「眾生喧嘩」的觀點，主張在小說寫作過程中，讀者對於角色乃是「只聞其聲，不見其人」，因此作者的首要任務就是要完整地呈現角色的話語。作者應該盡力釋放他對角色的影響力，以便能夠完整地呈現角色的道德觀與意識型態。由於角色必須發佈能夠代表自己社會、經濟、政治等族群的「語言」，不能只是做為作者意旨的傳聲筒，因此在文本上，讀者可以看到「作者話語」與「角色話語」兩者共存的「雙聲敘述」。 巴赫金的理論固然影響深遠，另一方面卻也備受質疑。許多知名的小說家與文學評論家如美國的那博可夫、亨利．詹母斯、英國的E.M.福斯特、法國新小說學派發起人葛力葉等，都認為作者對於其所寫作小說的重要性不可以被磨滅或削減。本論文認為，巴赫金理論的弱點在於他忽視作家對於小說架構全面性的設計與安排，也就是說，巴赫金所檢視的作者話語，其實不足以完整地驗證作者對於角色自主性話語的態度。本論文將巴赫金所述稱之「顯性作者話語」，本論文並補足巴赫金主張不足之處，另增加「隱性作者話語」作為理論分析的架構。簡要來說，「顯性作者話語」指文本中所有讀得到的作者的敘述；「隱性作者話語」則指作者對於小說的主旨、敘述方式、強化與削弱的層面、主要／次要角色的配置、小說涵蓋的時間與空間、小說的情節安排等在文本中雖然讀不到，卻深深紮根在小說的架構，紮紮實實地影響一本小說風貌的「作者話語」。巴赫金無視「隱藏性作者話語」的存在，使他忽略了小說家在寫作過程當中不可或缺的文學創作技巧，揭露巴赫金理論的缺點，也同時提醒我們後現代文學批評的一些盲點，特別是「作者之死」所引發的風潮，我們應由小說創作的實際過程，來重新思考該觀點的適切性。 M. M. Bakhtin's prepositions for the “polyphonic novel” have been widely influential as they are associated with several important concepts in postmodern literary criticism. Bakhtin believes that the more the authorial discourse seeks to engage in a dialogue with the character based on an equal position between the two, the richer the portrayal of the hero will be. He suggests that the author allow sufficient freedom for the hero to create his pure authentic voice. Nevertheless, the notion of the author cannot be easily annihilated, since it strikes its roots in every aspect of a novel, which surely brings a profound influence on the style of a novel. What makes Bakhtin's “polyphonic” novel questionable is because he concentrates only on the explicit authorial discourse, or the real wordy interactions between the author and his characters. He ignores the author's overall scheme of a literary work, which is mostly manifested in the latent authorial discourse. This paper intends to point out several weaknesses concerning Bakhtin's polyphony, together with my revisions of his theoretical statements. Diverging from Bakhtin's terms of analysis, I divide the authorical discourse into two categories--“visible authorial discourse” and “invisible authorial discourse.” I suggest that we add the analysis of the “invisible authorial discourse” to the discussions of novelistic language because it covers all elements on the structural plane of a novel and can express the author's intention in a more complete way. By carefully examining both “invisible” and “visible” authorial discourses, the reader is able to glimpse in completeness the author's willingness to allow the full development of the character's consciousness, ideology, and worldview. By combining the analyses of both “invisible' and “visible' authorial discourses, the reader can decide if a novel is really a polyphonic one.
