In “The Death of the Author,” first published in 1968, French literary critic and communication theorist Roland Barthes poses a fundamental question about the nature of any literary work of art: Who or what is behind it? Most readers would typically answer “the author,” “the writer,” or perhaps “the narrator” or the characters who speak. Barthes’s position, however, is that the question is unanswerable. A piece of writing does not contain any reliable record of the intention(s) of what readers commonly think of as an “author”—not what he or she meant, felt, or was trying to express. In seven short but demanding sections, the essay attempts to explain this counterintuitive thesis and some of the implications that follow from it.

This summary and analysis will refer to the seven sections numerically. They correspond to paragraphs but should not be thought of as typical paragraphs, as explained in the analysis. This study guide references the 1977 translation by Stephen Heath that appears in Image, Music, Text, published by Hill and Wang.

Barthes begins Section 1 by citing a passage from the story Sarrasine, written by 19th-century writer Honoré de Balzac. In the story, the title character falls in love with a woman he hears performing at the opera. However, the singer is not a woman but a castrato costumed as a woman. (A castrato was an adult male who had been castrated before puberty to sing in the high-pitched vocal registers of a woman.) Sarrasine is at first unaware the singer is a castrato, but even after he is told that “she” is not a woman, he doubts or is unwilling to believe that is the case.

Barthes quotes a passage in which the narrator expresses how the singer appears to embody a list of essential qualities of womanliness, such as impetuousness and delicacy of feeling. Barthes then asks his readers who is speaking this line. Is it Sarrasine? Or is it Balzac the ordinary man, drawing on his own lived experience? Or is it Balzac the professional writer, giving voice to characteristic French biases about women current in his day? Is it Balzac’s narrator, channeling some supposedly all-knowing source of wisdom? Barthes’s answer is that in this and in all literary writing, it is impossible to say who is speaking. He attributes this to the very nature of writing. In a famous statement, Barthes makes the startling claim that “writing is the destruction of every voice, of every point of origin” (142).

The word “writing” (écriture in French) has a special, unusual meaning for Barthes. The beginning of Section 2 helps readers understand this usage, although Barthes does not provide a straightforward definition. By “writing,” Barthes means language not when it is used to carry out a target objective in the world (for example, a cake recipe in a cookbook), but when it is practiced for its own sake, as in a fictional narrative or poem (though as he will make clear, there are other instances). The idea that communication of this sort has no identifiable “author” is not new, he continues. It was something obvious in more primitive cultures, where, say, a shaman might perform or channel a story or myth whose source is understood by hearers to be obscure or unknowable.

The modern world, however, which at least since the late Middle Ages has emphasized the primacy of the individual person, has accorded “the author” an exalted status. Critics and readers have regularly but mistakenly obsessed over the idea that understanding the meaning of any work of art requires knowing the creator’s essential biographical facts and traits, which serve as clues to his or her intentions. Some writers have even subscribed to this notion. Barthes hints that his thesis about the death of the author applies not only to literary writing (he mentions the poet Charles Baudelaire) but also to other and possibly all forms of artistic creation. He mentions the Dutch painter Vincent Van Gogh and Russian composer Peter Tchaikovsky as artists whose work critics have sought to explain through their biographies.

In Section 3, Barthes surveys some modern writers who nevertheless grasped this specialized understanding of “writing.” The 19th-century French symbolist poet Stéphane Mallarmé was the first to fully understand that “it is language that speaks, not the author” (143). Another symbolist, Paul Valéry, held the same view and regarded any concern with the author’s intention as “superstition.” The French novelist Marcel Proust is another example; his writing was intent on blurring any distinction between himself and his characters. By making his epic novel (In Search of Lost Time) a prolonged account of the inner psychological experiences of a character who is “going to write” the very novel he is in (Barthes’s italicizing of these words is key), Proust demonstrates his understanding that “writing” is fully disconnected from the author.

Surrealist artists—who sought to use the unwritten codes and conventions of art against itself, frustrating the audience’s desire to arrive at some distinct meaning in a work of art—also worked to “desacralize” the primacy of the author. Barthes mentions some surrealist writers’ experimental practice of “automatic writing,” a method of attempting to produce a text without intrusion from the writer’s conscious intentions. In the same vein, contemporary linguistics experts have sought to explain how communication between individuals involves operations of language that have nothing to do with the communicators’ identities or conscious intentions.

Section 4 focuses on these advances in linguistic science, examining how they support Barthes’s argument about authorless “writing” and some of the implications that must follow. One implication is that time as we commonly conceive it is radically altered: If there is no author, there is no act that “precedes” the text, nor any act of reading that “follows” it. The disconnection between the initiator and receiver of a piece of communication is not simply because there is a time lapse between one person employing spoken or written language and then another person attempting to decipher a message and meaning upon encountering those words. That linear, time-bound way of understanding communication still encourages us to think of an “author” as someone who came before the piece of writing “as a father to his child” (145). Barthes here introduces the term “scriptor” to replace the term “author.” Unlike an author or writer in conventional understanding, a “scriptor” is operated by language as it is recorded, somewhat in the way that a scribe takes down or copies a preprovided message according to rules of grammar and meaning that have nothing to do the scribe’s intentions. Paradoxically, it is better to say that the “scriptor” is created by and in the text itself, not the other way around.

An important consequence of this is that a text has no history; it is always only present in the here-and-now act of reading. “Writing” (écriture) belongs to a special class of communication acts that linguists call “performative.” These are utterances wherein the meaning of what is said is identical to and simultaneous with the action it carries out, such as in phrases like “I promise to visit tomorrow,” “I hereby deputize you” or “With this ring I thee wed.” The origin of any text lies not in its author but in the authority of language itself, whose resources the scriptor merely draws upon. (Thus “text” is another word that has a special meaning for Barthes.) The modern scriptor cannot entertain the idea that they are engaged in a singular, laborious, and unique act of personal creativity that was once supposed to be the business of “the author.” The scriptor is only the vehicle for a “pure gesture of inscription,” tracing “a field without origin” beyond even that instance of language itself that appears as a piece of writing (146). From here, Barthes lands at a large and consequential entailment of his argument: Every instance of writing in this sense “calls into question all origins” (146).

In Section 5, Barthes suggests further implications of “writing” (in his special sense) that go beyond our understanding of literature and art, and that are concerned with this obscurity of “all origins.” If every instance of writing in this sense renders uncertain our understanding of the author and the relation between the supposed author and the text, then this must include the idea of a divine creator (an “Author-God”) whose definite “theological” meaning can supposedly be extracted or derived from a text. No text can be said emanate from some singular, reliable, oracular message-sender; instead, readers must understand every text as a site where numerous voices are engaged and mixed into a “tissue [or fabric] of quotations drawn from innumerable centres of culture” (146).

In this sense, readers should recognize that in addition to the word “writing,” Barthes has been using the word “text” according to this specialized meaning—not simply a singly authored publication but a complex fabric pulling together multiple voices and influences. Although Barthes does not explicitly mention the Bible here (which is itself made up of many different books, each written at a different time by different and often multiple writers), he clearly gestures toward it. Barthes also mentions an unfinished novel by Gustave Flaubert as the perfect allegory for the predicament of the “text.” Bouvard and Pécuchet is the story of two copy clerks who go on a farcical quest to get to the bottom of all human knowledge (i.e., to become like gods). Likewise, Barthes implies that the reader who seeks the original author or fixed meaning of a text is embarked on the same hopeless and “comic” undertaking.

Barthes continues that the so-called author may think that they are expressing some original inner thoughts and/or emotions, but writing is no more than the evidence of a scriptor going through motions that are preordained in language, with its arbitrary but formally established words and grammatical conventions. Any final interpretation of a text is unreachable. Any understanding of a text’s meaning must be “infinitely deferred,” because, like looking up a definition in a dictionary, interpreting the written word can only be done by recourse to other words, and those words can only be interpreted by turning to still others, and so on without end. The scriptor can combine and mix elements of language but never truly “express himself.” Every attempt to express the self in that way only results in a “dictionary” of elements that are always already supplied in language and culture.

Here Barthes cites the French poet Charles Baudelaire’s description of Thomas De Quincey, a 19th-century English writer (famous as the author of the first addiction memoir, Confessions of an English Opium Eater, published in 1821). De Quincey recounts how, as a boy, he was so good at ancient Greek that he could translate any modern situation or idea into that old and “dead” language. According to Baudelaire, this resulted in a “dictionary” or repertoire of English phrasing that was far vaster than De Quincey could have developed by himself. In a way, ancient Greek “wrote” Thomas De Quincey’s texts. Thus, instead of thinking that art imitates life, as is commonly supposed, it is more accurate to say that “life never does more than imitate the book, and the book itself is only a tissue of signs, an imitation that is lost, infinitely deferred” (147).

At the end of Section 5, Barthes mentions “signs”—things that stand for or point to other things or meanings. This is significant, as he was one of the foremost theorists of how signs and symbols work (the discipline known as semiotics.) He opens Section 6 by saying that to think of a text as something created by an “author” is to attempt to foreclose further interpretation forever, giving a text a “final” unalterable meaning. This is something the very nature of signs disallows. Where readers would expect to find the word “meaning,” here Barthes uses the word “signified,” a term from semiotics that highlights how signs have fluid meanings acquired through complex interrelation with one another.

Barthes then shifts attention to another important implication of his thesis. Once readers understand that a text is not a creation of an “author,” and that nothing they know about the author will reveal anything definitive about the text’s meaning, then the professional literary criticism of Barthes’s day is shown to be a sham. Critics seek to establish their own importance by insisting, based on their supposed knowledge of the author, that they can locate inside the text the author’s intention to express some thought or feeling, a final meaning they insist on nailing down. Barthes argues that the most a critic can do is trace out the various voices of language interacting and colliding with each other in a text, to show how “writing ceaselessly posits meaning ceaselessly to evaporate it” (147). The word “literature” should be dispensed with altogether and replaced with “writing” (écriture) to acknowledge that both the author and the critic’s assessment are mirages, cultural hallucinations that some divine force is at work behind the text. Once again, Barthes expands the limits of this claim to apply it to all disciplines that use writing, including law and science—fields that people often “worship” as substitute gods because they imagine their authority to be fixed, defined, and tied to the intentions of a mythical author.

Barthes’s seventh and final section briefly circles back to the quote from Sarrasine. As soon as he returns to it, however, he leaves it to focus on an entity that has been lurking in his essay to this point: the reader. Reading is the only “source” in writing. (Note that Barthes first says “reading,” not “the reader.”) In support of this, Barthes points to then-recent research showing that ancient Greek tragic drama had a similar ambiguity at its center. The very nature of tragedy is the characters’ inability to understand one another even while using the same language. Only the spectating audience understands what the characters themselves do not, which is that the characters completely misunderstand one another. The audience “hears the very deafness of the characters speaking” (148).

Barthes is quick to say, however, that this reader is no particular person (nor any group or class of persons), only an anonymous “someone” on whom the text is written, a mysterious entity who “holds together” all the disparate voices that comprise the text. Here Barthes reveals how radical his claim about writing truly is. He imagines critics, probably outraged by his claim about the death of the author and the phoniness of their trade, racing hypocritically in defense of an equally almighty reader who has the “right” to determine and fix the meaning of a text. Readers must no longer be “fooled” into believing in such a reversal in their own favor, however. That is the very kind of thing that “good society,” still in thrall to the importance of the individual, might be happy with. The old notion of the Author, however, already “destroys” that kind of a reader as a final locus of meaning for the text, so any critic’s defense of the reader ought not be believed.