The Linguistic Illusion of Alterity: The Free Indirect as Paradigm of Discourse Representation

Abstract

In the winter 1991 issue of Diacritics, the two leading articles trace the reverberations of a long-standing debate between the adherents of a formalist and, as they claim, truly linguistic approach to language and the proponents of a pragmatic, antiformalist conception of language. Oswald Ducrot's contribution discusses the importance of Charles Bally as an unrecognized precursor of latter-day pragmatics, particularly speech-act theory; Ann Banfield, in turn, attempts to annihilate pragmatics from the perspective of ruleoriented, empirical linguistics. The ostensive topic of the two articles is the status of voice within a polyphonic account of linguistic expression. Ducrot rewrites Bally's distinction between the speaking subject and the modal subject in the light of his own theory of linguistic enunciation, proposing a tripartite distinction between a communicating subject (in the first example sentence of our epigraph, the gentleman), a speaking subject (the maid who utters this sentence to her employer), and a modal subject (the gentleman's wish to speak with Madame) [12]. Likewise, in La Fontaine's famous characterization of the miser's fear of burglary (our second epigraph) Ducrot notes a noncoincidence between communicating [the narrator] and modal [the miser] subjects, arguing that La Fontaine' s line combines the two points of view in a single utterance [17]. Ann Banfield, whose paper is a response to Ducrot's criticism of her work in his book Le dire et le dit, characterizes Ducrot's proliferation of voices as a-to her mind-illicit multiplication of speaker functions and maintains that her own seemingly more unitary, univocal theory of discourse is the more radically polyvocal proposal. Banfield's theory dispenses with the unitary speaker function and, in an adumbration of polyvocality, allows texts to shift from one SELF (center of consciousness) to another at each sentence (S) boundary. The alternatives between Ducrot's model and her own, she argues, therefore need to be