The voice you distrust

**The term was coined in 1961 and spent the decades since being argued with, including by its author.** Booth's definition is one line and it decides by reference to somebody who is never on the page: "I have called a narrator reliable when he speaks for or acts in accordance with the norms of the work (that is, the implied author's norms), unreliable when he does not." The reliable distinction is therefore not a property the reader can check in the story; it is a verdict reconstructed from the shape of the telling, by a reader standing alone with it. The essay read for this entry opens by recording the distance the term has travelled - sixty years have passed since Wayne C. Booth conceptualized narrative unreliability - and traces Booth's own 1983 return to it as well as the expansions the concept has received since, which is the honest state of the word: coined once, revised by its coiner, extended by everyone else. The corpus keeps the coinage and its date, because the year is checkable and the boundaries are not. **Why it belongs on this site rather than in a craft manual: the reliability of a narrator is maintained entirely by the reader, which is this site's structure applied to a voice.** No character in a story can be persuaded to reconsider; no implied author can be reached for clarification; and as the fictional-protagonist shelf's mechanism has it, "because the text never closes the account the audience is never caught being wrong" - the unreliable case is that mechanism with the volume up, because now the audience is not merely filling in an interior but auditing one, from outside, forever. The corpus's instance is fight-club, whose Narrator the entry quotes directly: "The Narrator, an unreliable narrator, is not immediately aware that he is mentally projecting Tyler" - a speaker whose account is the only account there is, contradicted by nobody, believed by the reader for two hours on the strength of nothing but the telling. Booth's own 1961 framing suits this reading: the judgement falls to the implied author, and the implied author is a thing the reader builds to have somebody to trust. **The counter-argument is that unreliability is a property of texts rather than of relationships, and that filing it here confuses a narrative device with a bond.** It is the strongest objection available and it is half right: there is no far side scheming to be believed, and every other entry on this site has at least a figure on the other end. It loses on what the reader is actually doing. Distrust spent on a narrator has the same anatomy as every tie this site files - attention paid unilaterally, an account supplied without the far side's knowledge or consent, and a correction that never arrives - and the term itself names the reader's half of it rather than the writer's. The fourth-wall entry, dated to Diderot in 1758, describes the arrangement from the outside: the wall is the fiction's silence. The unreliable narrator is the voice inside that silence the listener cannot audit, and the audience's response is the same in both cases - it keeps attending. Filed beside fight-club on the written-character shelf, where the account that never closes is the shelf's own mechanism; the counter `character/played-role` loses because the doubt here is aimed at a narrator rather than at a performer, and the reader has no actor to hold. Contestable, and recorded as such.

Application. One-sided attachment put to work on purpose. Creator economies, community, fandom as infrastructure, and the things that genuinely help people. Not automatically good - a tool with an obvious use is a tool that will also get a bad one.

Consideration. The part where somebody stops and thinks. The research, the diagnosis, the ethics, the lawsuits, the regulators. Roughly fifty years of arguing with itself about whether this is a relationship.