[Access article in PDF]
The Distinction of Fiction
The Distinction of Fiction, by Dorrit Cohn; ix & 197 pp. Baltimore: Johns Hopkins University Press, 1999, $42.00.
In this wonderful book, which expands and revises several of the author's previously published papers, Dorrit Cohn underlines the uniqueness of fiction by examining formal patterns that only fiction can accommodate. Resisting the tendency to erase boundaries between different orders of discourse and, in particular, between fictional and nonfictional narrative domains, Cohn emphasizes and explores those boundaries by focusing alternately on theoretical problems and specific texts.
After reviewing the various meanings attached to the word "fiction"--fiction as untruth, fiction as conceptual abstraction, fiction as (all) literature, and fiction as (all) narrative" (p. 2)--and after arguing that, in a literary-critical context, the word should designate only nonreferential narrative texts, Cohn specifies a number of criteria distinguishing fictional narratives that focus on the lives of individuals from historical biographies and autobiographies (for example, the representation of consciousness in third-person novels or the difference between author and narrator in first-person novels). To show that borderline cases highlight the dissimilarity between fiction and nonfiction, Cohn studies Freud's most famous case histories (Freud is simply not a novelist), Proust's A la recherche du temps perdu and the sources of its generic ambiguity, Wolfgang Hildesheimer's Marbot (a fictional biography adopting the narrative stance of historical writing), and first-person narratives like J. M. Coetzee's Waiting for the Barbarians, which use the present tense from beginning to end and cannot be subsumed under established narrative categories or be made to correspond to natural discourse (the present of these "simultaneous narrations" is neither the historical present nor the present found in interior monologues). Instance after instance reinforces the validity of Cohn's separatist thesis. Indeed, there are signposts of fictionality and Cohn devotes an important chapter to identifying and discussing three of them. First, fiction can be narratologically characterized in terms of a synchronic, bi-level story-discourse model that is independent from a previously constituted referential data base. Nonfiction, on the other hand, must be characterized in terms of a diachronic, tri-level model--reference, story, discourse--in which the interaction of the latter two levels depends on the logico-temporal priority of events in the real world. Fictional narratives are plotted whereas nonfictional narratives are emplotted; and generic borderline cases like Broch's Death of Virgil or Yourcenar's Hadrian's Memoirs help to illuminate the difference. Second, as mentioned earlier and contrary to nonfiction, fiction routinely uses narrative situations that allow access to the inner life of a character. Third, the narrator in both homo- and heterodiegetic fictional narratives is detachable from an authorial origin while in nonfictional narratives author and narrator are taken to be identical. Thomas Mann's Death in Venice provides Cohn with a paradigmatic [End Page 234] example of fiction in which, far from being the mouthpiece of the author, the narrator can be considered unreliable in the Boothian sense. Similarly, Tolstoy's War and Peace allows her to show that the historical novel differs most sharply from historiography through its use of internal focalization, its recourse to "inside views," its depiction of the characters' raw experiences. In a final chapter, Cohn warns against correlating manner and matter, technique and value, textual form and ideological orientation without taking the complexity of fictional devices into account. In linking narrative omniscience and panoptic vision, for example, critics like D. A. Miller in The Novel and the Police are unconvincing, to say the least. After all, narrative omniscience gives access to invisible minds whereas panopticism pertains to observable behavior. More generally, the relationship between narrative mode (or other narrative categories) and meaning varies considerably. Omniscient narrators are not necessarily reliable. Omniscient narrators do not all promote the same ideology. An internally focalized novel is not incapable of supporting unequivocally a particular political, social, or cultural agenda. If narratology is to be useful to sociopolitical examinations of fictional texts--and it can be--the distinction of fiction...