Few writers have imparted their views on questions about writing and authorship in as many genres and modes as Margaret Atwood has over the past half century. She has presented her ideas in reviews and interviews, introductions and afterwords, poems and lectures, novels and short stories, and hand-drawn illustrations (which can also be found on t-shirts). Even her invention the LongPen, a device for providing handwritten dedications from remote locations, has been interpreted as a comment on authorship (Wolframe 13). Indeed, and only partly because they have been presented in so many contexts, her views on authorship sometimes contradict one another. My sense is that this is not a shift over time or simply that different contexts call for different takes on authorship. Instead, I see a wavering that indicates a third option, only sometimes articulated. In certain texts, Atwood seems to be asserting the existence of an independent author, positing what may even be a Romantic vision of authorship. At other times, the author seems to be merely a channel or instrument through which somebody else’s voice flows. In both cases, Atwood often implicitly subverts the view that she presents explicitly. After giving some account of these two points of view, I will describe a middle ground between them through a close reading of Lady Oracle (1976), the novel that I find has the most material for thinking through these positions.

This split within Atwood’s oeuvre corresponds to a split within literary studies. On the one side, we find text-centred approaches — such as poststructuralism — that see writers as conduits for forces or systems beyond their control, such as language or power. In this approach, readers should ignore the author. On the other side, we have the incessant practice in academic criticism (as well as literary journalism) of organizing research around authors and their lives. Furthermore, some theorists, such as James Phelan, argue that ignoring the author and what she intended to convey narrows the field of literary inquiry (131-33). This debate is often framed by the question of how much critics should investigate and take into account authorial intention, but it can have another emphasis: the extent to which authors can be said to be the originators of their works. Several critics have understood Atwood to be supporting and enacting the poststructuralist view of the writer. As I will explain, Bruno Latour’s ideas and their recent adaptation to literary studies, especially by Rita Felski, allow a substantial middle ground between the author as irrelevant conduit and the author as key originator. [1] In adapting Latour’s and Felski’s thinking to the discussion of literary composition, I will show how Atwood often presents authors as mediators, messengers who change messages in significant ways. In reading Atwood, I will also highlight how Felski’s framework can better incorporate the author and other agents who are part of the network of literature but precede the publication of the text.

In doing so, I will attempt to continue in my own way Lorraine York’s project in Margaret Atwood and the Labour of Literary Celebrity. York does not engage directly with Latour or Felski, but she minutely describes the processes that create not only Atwood’s celebrity (as the title suggests) but also the literary texts that comprise the most solid foundation of her fame. York focuses for the most part on how this process works in reality and not on how it is depicted in Atwood’s writing; however, Atwood’s work calls for exploration of the latter kind. Unlike Felski, who tends to focus on readers and reception, York offers a focus on producers. The need to respond to Felski’s intervention along with the perennial relevance of authorship to literary studies makes it important to reopen the questions of authorship now.

Perhaps the most well-known and often taught iteration of the post-structuralist framing of authorship is Roland Barthes’s “The Death of the Author.” [2] In this essay, Barthes argues against the picture of the author as an originator of meaning, someone who expresses himself through his writing. Reading, according to Barthes, should not be a search for what the author meant to communicate or the expression of his personality, biography, or even social position or historical moment. In Barthes’s essay, the writer (“le scripteur,” as he calls him; see “Mort” 43; “Death” 145) is only a hand through which words flow, perhaps through his assembly but not in a way that expresses his unique personality. In fact, for Barthes, the scriptor does not exist prior to the writing but is born with it (“naît en même temps que son texte”; see “Mort” 43; “Death” 145).

Barthes stresses that the author is a modern invention: “Dans les sociétés ethnographiques, le récit n’est jamais pris en charge par une personne, mais par un médiateur, shaman ou récitant, dont on peut à la rigueur admirer la ‘performance’ . . . mais jamais le ‘génie’” (“Mort” 40). [3] The “genius,” the source, is someone else, often otherworldly. In premodern societies, then, the storyteller was free from the cult of personality in which the modern author finds herself, and Barthes hints that a return to an older model is part of what he has in mind when he proclaims the death of the author. In “Descent,” the final chapter of Atwood’s 2002 non-fiction book Negotiating with the Dead: A Writer on Writing, Atwood stages such a return in a way that makes it tempting to align her with Barthes. There she implies that authors merely transport literary material from one realm of existence to another, that they are mediums between the dead and the living. She presents the “hypothesis . . . that not just some, but all writing of the narrative kind, and perhaps all writing, is motivated . . . by a desire to make the risky trip to the Underworld, and to bring something or someone back from the dead” (156). Here the writer is a traveller to the Underworld, where he or she interacts with the departed. Writers seem to be messengers to the living: “The dead may guard the treasure but it’s useless treasure unless it can be brought back into the land of the living . . . [and] enter the realm of the audience, the realm of the readers” (178-79). Despite the empowering language of shamanism and inspiration, the writer is merely a conduit of knowledge-treasure from a source to its destination.

The idea of the writer as a medium — an instrument for passing language from a different realm of existence, a scribe who takes down the words of another agent — is an ancient one. Harvey Hix, who presents a thought-provoking taxonomy of modes of authorship, calls this role “the transcendental mode,” and he identifies it with the writing of some parts of the Bible, in which a scribe takes down the word of God, and with Greek oracular practices, in which the oracles and other bearers of signs were the modes of communicating with the gods (90-96). Hix points to the Ouija board as one of the main incarnations of this mode in modern society. Unlike Atwood, he does not go into other Spiritualist-influenced practices (92). In all of these practices, the authority of the text comes from a transcendental sphere, not from the individual putting pen to paper (or parchment). Many, but certainly not all, of Atwood’s sources in the chapter are classical or medieval, going back as far as the Epic of Gilgamesh and including Virgil and Dante. However, Atwood is making an argument not about classical, medieval, or non-Western cultures but about writing in general. She claims that this mode is relevant for authors today.

Atwood is far from the only contemporary author, though, to suggest that they or authors in general are mediums and therefore not in control of their texts. One example comes from the authority of not only one but two Nobel laureates: J.M. Coetzee and Czeslaw Milosz. In the final chapter of Coetzee’s novel Elizabeth Costello, the eponymous protagonist — an older Australian novelist — finds herself in a Kafkaesque purgatory. There she is charged by a tribunal to write what she believes. She says that she has no beliefs because “I am a writer and what I write is what I hear. I am a secretary of the invisible. . . . That is my calling: dictation secretary” (199). She is never asked about the voices that she hears, but they come from “powers beyond us” (200). As she also says, the phrase “secretary of the invisible” is taken from “Secretaries,” a poem by Czeslaw Milosz (325), which voices much the same approach to the task of the writer. Coetzee, unlike Atwood in Negotiating, distances himself from this opinion by attributing these views to a fictional character (who might herself be misrepresenting or simplifying her outlook). Still, Coetzee cares enough about this concept of authorship as supernatural secretarial work to put it forward in his fiction.

This conceptualization of authorship seems to be surprising because the idea of the author as a messenger does not easily tally up with modern ideas about individual authorship and creativity, concepts consolidated in the Romantic period and upheld ever since by copyright laws, which ensure ownership of texts by the individuals who wrote them. [4] Yet the idea of the author as medium seems to be supported by Barthes’s “Death of the Author.” Indeed, PhebeAnn Wolframe relates a series of Atwood’s texts to Barthes’s formulations. For instance, Wolframe responds to an epigram that Atwood keeps pinned on her office’s bul- letin board — “Wanting to meet an author because you like his work is like wanting to meet a duck because you like pâté” — and Atwood’s interpretation of it (Negotiating 35). Wolframe contextualizes this quip within an overall sense that for Atwood the author as person disappears when the writing is done. Wolframe writes that, “at least metaphorically,” “[t]he Author, the one who writes, can never be present, because the Author, like the duck in the paté [sic], is dead,” and she goes on to quote “The Death of the Author” (19). I think that Wolframe is too quick to bridge the gap between Atwood’s point, based upon the notion that authors are not really like their books, and Barthes’s stronger claim that they do not originate their books. As for the death of the duck, perhaps Atwood is invoking the powerful title “The Death of the Author,” but she does so ironically. As metaphors, Atwood’s statements are in sync with “The Death of the Author.” However, the humour and irony in her depiction of the downward trip make one question the extent to which Atwood unambiguously supports the idea that authors are conduits for others.

In other moments, Atwood gives the impression that authors are the origins, even the sole origins, of their texts. For her, the author is sometimes the genius that Barthes was trying to dethrone. In a 2011 talk that was part of a tech-oriented conference, Atwood insists on the primacy of the author in the ecosystem of literary production and consumption. She warns publishers and technologists: “Never eliminate your Primary Source” (“Publishing” 8:30). Thus, she places the author at the centre of textual production by presenting two of her illustrations of a dead moose as the primary source of food for a variety of animals and then a parallel one of a dead author who feeds publishing and education. Yet the author here is not only alive but also makes it possible for others to live and write.

But, much like the irony in Negotiating, I suspect that Atwood embeds a counter-argument within the argument she explicitly makes. Atwood presents the author as a starting point and not as a conduit, though she could have shown how the author also feeds off other sources. Lorraine York has shown in detail how much labour by other people such as “editors, agents, office staff, publishers, publicists, and the like” goes into creating not only Atwood’s texts but also her public persona (8). We can also think of the author as feeding off other authors or, in line with Atwood’s other writings, the dead. This counter-narra- tive of the author as feeding off others is visible if we think a little more about the moose analogy. Moose eat grass, lichen, and other organisms that photosynthesize, making energy from the sun available to life on Earth. A moose is not the primary source, but this is not how Atwood chose to present the situation. Perhaps, since she knows a fair deal about ecology, she chose this analogy to hint that the image of the author as the centre of the literary system is not the only story that one could tell but the most convenient one for the purpose at hand.

Atwood also places authors in a primary position in the introduction to Second Words: Selected Critical Prose (1982). She explains the choice of title with the common-sense notion of “precedence: that is, a writer has to write something before a critic can criticize it” (11). Here she regards authors as the source of their literary works. When, in her 1976 essay “On Being a ‘Woman Writer’,” she thinks that she is being asked to be a spokesperson for feminism, Atwood insists that “no good writer wants to be merely a transmitter of someone else’s ideology, no matter how fine that ideology may be” (203). A good writer, for her, would never just repeat ideas that do not belong to her. She would not be a “transmitter,” a medium, or a mere intermediary. Insisting on individuality, Atwood refers to writers not only as “eye-witnesses” but also “I-witnesses” (203). The author is, first of all, an I.

This view is also represented in Atwood’s fiction, for instance in the short story “Alphinland” (2014), which presents a fantasy fiction writer named Constance. Like much of Atwood’s fiction, the story goes back and forth between present and past. In the present, Constance is a recent widow facing a blizzard, while her memories of the past focus mainly on her time in the Toronto bohemia of the 1950s. During that period, she was a lover and muse of Gavin, a poet. She also started writing fantasy stories about Alphinland. Working initially to make money (to support the supposedly true artist, Gavin), she nonetheless ends up using her fiction writing as a way of asserting herself: “Alphinland was hers alone. It was her refuge, it was her stronghold” (Stone Mattress 22). Writing stems from the author, described as “its creator, its puppet mistress, its determining Fate” (17), words indicating the opposite of what Barthes claims of the “scriptor”.” Constance is the only one who possesses this place; her deceased husband could never enter it. He once drew a gateway to Alphinland for her but could not create a scene from the land itself, merely the gate to it, the image of his exclusion. Although he can imagine the gateway, he still “misses the point” (16). In this story, even seemingly commercial or genre writing becomes a place for self-expression and self-reliance. It is an opportunity for a romanticization of authorship.

Some sense of individuality is presented through the fantasy world, which contains Constance’s biography. Atwood relates it not in the usual manner of presenting some fictional characters or plotlines based upon real people and events, but by imagining people whom Constance knew trapped inside the story world: “Thus she keeps a deserted winery in Alphinland . . . for the sole purpose of Gavin. . . . [H]e’s preserved in a state of suspended animation. Every once in a while . . . [Constance] says a charm that unlocks the top of the cask and has a look” (Stone Mattress 18-19). Alphinland is where she keeps her personal history, and though it is not stated explicitly, readers are led to understand that it somehow powers the endeavour of writing. Alphinland and the texts that describe it are the expression of Constance as a unique individual. At the same time, Atwood shows some discomfort with this way of thinking, exemplified when Alphinland is described as if it is a place that actually exists and that Constance merely explores (and does not build by writing). Thus, “Even though she’s its creator . . . Constance never knows exactly where she might end up” (17). The possibility that she is just relaying material from the alternative world is opened, even if it is never supported by this story itself. [5]

It is clear, then, that Atwood rarely insists on authorship as either primary or secondary. However, her third novel, Lady Oracle (1976), engages with this pair of ideas through communication with the dead in a way that creates a middle ground between the genius and the medium. The novel shows that authors are neither sole original creators nor passive conveyors of communications from others. Rather, they are mediators who receive material and change it according to their own personalities and agendas.

Latour and Felski give us some tools to examine this middle ground at the same time that Atwood’s thinking about authorship expands Felski’s intervention in critical practice. I agree that critics should indeed think of authors as links in the network of literature. However, as Latour helps us to see, links in social chains are crucial and can create change. I will make two interconnected points through his argument. The first is that authors are, often think of themselves as, and present themselves as, links in the chain of literary assemblage (or as it is sometimes called, composition). The second is that as links in a chain, they are still mediators rather than intermediaries. That is, they are neither sole originators nor irrelevant conduits of textual material. The necessity of such an adjustment becomes evident through reading how authors describe the process of composition. Through the depiction of an author-character in Lady Oracle, Atwood shows how authors are nodes in the network of communication and as such make a difference.

I take this set of terms — “mediator” and “intermediary” — from Latour’s Reassembling the Social: An Introduction to Actor-Network-Theory, in which he demands that social researchers refrain from seeing non-humans (animals, things, concepts) as merely intermediaries and begin asking instead when they are mediators that shape social interactions. For example, even if we do not always notice them, telephones and telephone lines do not merely facilitate social interactions among people but also shape their interactions and are parts of them. For Latour, they too should be conceived of as mediators. The mediating nature of a telephone is clearer when we think of a conversation on a malfunctioning one. The way that we are at the mercy of mediators becomes almost visceral when one is yelling “What? You are breaking up,” becoming angry at the human interlocutor and only later realizing that the telephone set the tone of the conversation. Yet we should remember that both faulty and functioning devices are part of any interaction that the existence of telecommunications enables.

On the face of it, literary critics do not need a reminder that writers are not intermediaries. However, when theorists, critics, and even authors radically reduce the role of the author in the network of literature, such a reminder might be useful. Felski argues that texts should be treated as actors, which also means that they are “mediators” (the Latourian term that I find more useful, especially for a discussion of a text about a spirit medium). For Felski, texts are “not actors in [a] rugged, individualist sense. . . . If they make a difference” — that is, if they mediate — “they do so only as co-actors and codependents, enmeshed in a motley array of attachments and associations” (“Context” 589). Literature is a web, part of a larger social web, that enables texts to connect and act on other actors, including, most importantly for Felski, readers. She does not explicitly include (or exclude) authors in this network of association. This seems to me to be an unfortunate omission since, as she puts it, if “Artworks can only survive and thrive by making friends [and] creating allies” (584), then surely some of these friends and allies are the actors who assembled the texts in the first place, among them the author. Building upon Latour’s work means that to define anything as a mediator, something that helps to shape the interaction of which it is a part, it is necessary to look closely at how it functions. This method also justifies the relatively large proportion of this essay dedicated to a single aspect of a single novel, to which I now turn.

The author, then, is not dead, but she might seem so at times. Indeed, Lady Oracle’s main author-character, Joan Foster, tells her life’s story after having faked her death. She is an author who seems to be dead but is not. Whether Atwood was thinking about Barthes’s essay when writing Lady Oracle I cannot say for sure. However, death was an important way for her to think about authorship, just as it was for Barthes and the countless critics who have been influenced by his essay. Joan presents herself as a modern oracle, one who communicates not with the gods but with the dead. Seeking solutions to artistic problems and new sources of authority, she engages in spirit mediumship. This practice leads her to the creation of a commercially successful book of poems, which launches her career as a famous author. Simply put and in line with Negotiating with the Dead, Joan needs the dead in order to become an artist. The novel shows us — and this is not clear from the non-fiction book — that even when a writer brings a treasure from the dead it is no longer the same treasure; it is the author’s as much as the Underworld’s.

Her first encounter with Spiritualism comes when Joan is still a teenager and her Aunt Lou takes her to a Sunday service in a Spiritualist chapel. Spiritualism is “a religious movement emphasizing the belief in survival after death, a belief spiritualists claim based upon scientific proof, and upon communication with the surviving personalities of deceased human beings by means of mediumship” (Lucas 337). In Latour’s terms, it is a practice that promises mediation between two realms. The chapel is shabby, the practitioners elderly; nevertheless, Joan takes their beliefs to heart. One of the hymns from The Spiritualist Hymnbook is about “the blessed spirits . . . watching o’er us.” “This thought” of the dead “spying on” her makes Joan “uncomfortable” (Lady Oracle 106).

The leader of the church, a “stately older woman” who calls herself Leda Sprott (106), introduces Joan to the practice of Spiritualist automatic writing. After the incident in which Leda sees Joan’s mother’s astral projection, Leda encourages the teenaged girl to pursue such writing. In what seems like a parody of a scene in which an older writer encourages a younger one, Leda tells Joan, “You have great gifts . . . great powers. You should develop them. You should try the Automatic Writing, on Wednesdays” (112). [6] Her talent should be nurtured. Her gifts are to manifest themselves through the practice of writing. However, the addition of “the” before “Automatic” and of “on Wednesdays,” that is to say, at the chapel under her supervision, makes the encouragement seem like a sales pitch, deflating the pathos of the moment. Yet this promise of power excites Joan, and she fantasizes about becoming a famous medium. The parodic quality of the scene connects automatic writing to authorship even as it envisions both as labours of mediation.

The early experiment with automatic writing is hazardous and unrewarding, for Joan enters a trance but regains consciousness to find her hair on fire and no legible message. It is only when she is older, has lost close relatives (her mother and an aunt), and already has a career as a writer that she can produce whole words by automatic writing. Even then Joan does not undertake automatic writing in the hope of producing something new or artistic; she does not have a poem in mind. It is, rather, a way of solving a problem in the writing of popular fiction. Joan needs a change, and at first she believes that a small one will do. She decides to insert an element of the occult, now a popular device in the fiction of her rivals, into her own novel. Going to the library, she states that “What I needed was a ritual, a ceremony, something sinister but decorative” (217). The sentence overtly signifies that her Gothic novel requires this element. However, the exact phrasing — “What I needed” — hints that Joan requires the supernatural in her life to get over her creative malaise. She does not need the minor addition of weird decoration to her romances, but a more radical change in how she writes. She needs to connect to an alternative network. She actively searches for a new practice that eventually seems to reduce her agency. In the trance state, she seems to be the image of passivity and intermediality. It is important, therefore, to remember that she reached this state through a desire to find a new creative path.

Indeed, as an adult, Joan’s first attempt at automatic writing is brought on by a specific problem in writing a novel. She is stuck with a scene and therefore needs to act it out, as she has done in the past. The villains induce a trance in Penelope, the heroine, wishing to know what she sees and hears. Contrary to the case in Atwood’s novel, in Joan’s pulp romance other characters try to turn the protagonist into a medium. At this moment, Joan does not know how to continue the scene. This relatively minor setback encourages her experimentation with automatic writing. She tries automatic writing to solve a problem with production, not for any spiritual reason.

In attempting to imagine the scene of Penelope’s trance, Joan enters a trance of her own and feels that she is searching: “I was going to find someone. I needed to find someone”(220). She is looking for a source of automatic writing, an author who will write for her, a dictating voice. And, indeed, in her first adult trance, Joan seems to see the person for whom she is looking: “There was movement at the edge of the mirror. . . . Surely there was a figure standing behind me” (220). She finds an author, a source, to take over her hand and write for her. She seems to have been taken over by another agent, one akin to language or power in poststructuralist accounts. It seems as if she is just an intermediary. When she looks at the page, she finds “a single word: Bow” (220).

In her trance, Joan goes to a mythological underground realm of the dead. She describes her first experience as “walking along a corridor, I was descending” (220); later she experiences “a narrow passage that led downward” (221). Although she never arrives at a destination beyond the descending passages, the words that appear in the poems also indicate a netherworld, with terms such as “under the earth’s arch” and “death boat,” and the main figure of the poem, the lady, is said to “live under the earth somewhere” (222). Joan participates in what Atwood discerns, in Negotiating with the Dead, as the archetypal journey of the author to the Underworld. These images have also played a significant part in her poetry (see Huebener 110-13). She casts herself in the role of mediator, someone who can carry material from one site to another one.

Despite her fear that her behaviour is “too ridiculous,” Joan continues to employ automatic writing because of the high stakes that she attributes to it: “There was a sense of . . . the certainty that if I could only turn the next corner or the next . . . I would find the thing, the truth or word or person that was mine, that was waiting for me” (221). And, indeed, if Joan has a sense of purpose, part of it surely comes from the sender’s identity. This “person that was mine” is the author of words that become the bases of the poems; this figure embodies the destination and purpose of Joan’s travels in the underground world. Toward the end of the novel, Joan acknowledges almost explicitly what many readers would already have suspected: her mother’s spirit is the source. After dreaming about her mother, Joan thinks of how she has continued to haunt her: “She’d never really let go of me because I had never let her go. It had been she standing behind me in the mirror, she was the one who was waiting around each turn, her voice whispered the words. She had been the lady in the boat, the death barge, the tragic lady with the flowing hair and stricken eyes, the lady in the tower” (329-30).

Returning to the details of the automatic writing experience (the figure behind Joan, the turns in the imagined corridor), and the images and main character of the poems (the lady on the death barge), readers sense that Joan believes that her mother was responsible for the writing of these poems. The most explicit acknowledgement of her mother’s involvement comes when Joan states that “her voice whispered the words.” Although the poems are not explicitly mentioned, the context is clear enough to indicate that Joan believes that the originator, the author, of Lady Oracle is in fact her mother’s spirit. Joan knows that she has a role, but it seems to be only that of scribe or publisher. [7]

Let us return to the question of whether Joan is an intermediary or a mediator. At first, it seems that this process of automatic writing means that she is not the author of the Lady Oracle poems. If she is a medium, then she is, as the word suggests, nothing more than an instrument in the hands of her mother, used to communicate with the living. She is no more the author of the poems than the typist is the author of a dictated letter or, for that matter, the typewriter is the author of that letter (to give away some of the argument, in a Latourian analysis, both secretary and typewriter would be intermediaries). Indeed, several critics have described Joan’s position as all too passive. Mary Eagleton writes that Joan’s “writing is, then, a form of possession rather than authorising and Joan’s effectiveness as a writer is another manifestation of what Leda Sprott sees as Joan’s talent as a medium” (131). Fiona Tolan makes a similar point with the help of Barthes’s analysis of the mythology of authorship:

When Joan substitutes drugstore romances with epic poetry, she does so with a comic lack of the creative energy traditionally ascribed to the author’s task, which Barthes describes: “The Author is thought to nourish the book, which is to say that he exists before it, thinks, suffers, lives for it, is in the same relation of antecedence to his work as a father to his child” [Barthes, “Death” 145]. Abandoning this principle, Atwood’s narrator composes novels with her “eyes closed” and writes poetry in a self-induced trance. (65)

In Tolan’s argument, Joan suffered the Barthesian death of the author before she even became an author. Tolan, Eagleton, and others argue that Joan has little literary authority and that the procedure of her composition shows that she does not deserve credit for it.

I quote the critics who find Joan lacking in authority to make it clear that in the context of Lady Oracle the argument that she has authority to mediate is not a trivial one. In fact Joan does retain much of her power to mediate. A careful reading highlights how the procedure of writing the Lady Oracle poems is more complex than simply taking down dictations. An interviewer’s statement that “these poems were dictated to you by a spirit hand” is only half confirmed by Joan with a “Yes . . . something like that” (237-38). As she somewhat haltingly tells the interviewer, in the automatic writing “these words would sort of be given to me” (237; emphasis added). Words, not poems, are received from the “other side.” In carrying the words, she controls the poems. Even if some of their words come from a ghostly source, her mediation is part of the process of composition.

One of the best illustrations of this process occurs when Joan receives the first word. When she sees the word bow, she asks herself, “What the hell is that supposed to mean?” (220). She is aware that, outside of a context that she might give it, a single word can have countless meanings or no meaning at all. Indeed, when Joan looks up the word in Roget’s Thesaurus, she finds that it is not possible to know if it is a noun or verb and that it has various definitions. [8] “What a dumb word,” she thinks (220). Frustrated, Joan means that the word is stupid or vapid. However, another meaning of dumb is more accurate: the word bow, and words in general, are mute. Bow cannot express itself. Indeed, it cannot even vocalize itself: it can be pronounced in at least two different ways depending on its meaning. The word achieves articulation only when Joan puts it into a context and decides on its meaning. The word can be interpreted as a command to Joan to yield in obedience to a greater power, bending the knee, but it need not be interpreted thus. Instead, she decides that bow refers to the forward part of a boat. Joan might be deliberately giving the word a meaning that would work with a certain image of the spirit world that she already has in mind, one that features some version of Charon’s barge.

As Joan experiments further with automatic writing, her power over the poems becomes clearer. Here is another passage describing her creative process: “When I would emerge from the trance . . . there would usually be a word, sometimes several words, occasionally even a sentence, on the notepad in front of me. . . . I would stare at these words, trying to make sense of them; I would look them up in Roget’s Thesaurus, and most of the time, other words would fill in around them” (221). Even in her conscious state, Joan ascribes agency to words and not to herself. They would “fill in around” the automatic writing. Here it seems as though Joan is being used like an instrument by words. Yet it is clear that she consciously chooses words that will give meaning to the trance material. She is the one “making sense” out of these mysterious yet nonsensical communications from the Underworld. She is the originator of much of the text of each poem, the trance words acting as prompts for creativity. The ghostly presence that dictates words is not a poet but an inspiration. In many respects, Joan is no different from many other writers who describe loss of control over their manuscripts, with plots, characters, and words taking on lives of their own or works being discovered rather than invented. There is little difference between how Joan discovers words on the page and how contemporary horror novelist Stephen King describes starting work on one of his novels: “The actual story did not as then exist (well, it did, but as a relic buried. . . . I had located the fossil; the rest, I knew, would consist of careful excavation)” (167). [9] Even with no mention of the supernatural, an author can describe his creation as a found object.

The Victorian adventure novelist H. Rider Haggard also presented his work on the novel She (1886) as automatic. It is worth reading Haggard in this context for two reasons. First, She is alluded to when Joan includes the phrase “she who must be / obeyed” (226) in her poems; this is the immortal Ayesha’s full appellation, repeated many times in the novel that carries its first word as its title (see Haggard, She chapter IX). Second, Atwood was to study Haggard’s novels in her never completed doctoral dissertation, and she has continued thinking and writing about Haggard, most recently in a piece published in her collection of essays about speculative fiction In Other Worlds: SF and the Human Imagination (106-14). Haggard describes She as “written at white heat, almost without rest. . . . [I]t came faster than my poor aching hand could set it down ” (Days 10). Haggard is not suggesting anything like spiritual intervention or even tapping into an unconscious mind. Yet as Bruce Mazlish suggests, Haggard “seemed to write the book in a trance: a kind of ‘automatic writing’” (731). The text came to him, and his hand, discussed as if separate from him, is described as an insufficient intermediary between the unknown source and readers. Haggard, of course, had no qualms about presenting himself as the author of the novel.

Other critics have pointed out that in many respects Joan’s practices are not much different from those of other writers. Madeleine Davies, for one, argues that the poems are examples of Hélène Cixous’s écriture féminine (“women’s” or “feminine writing”). For Davies, “The most important aspect of Joan’s automatic writing lies in the fact that the language she produces sidesteps ‘the language of men and their grammar’ [Cixous 887] because it emanates from the unconscious rather than from the conscious mind” (66). [10] This byway to the unconscious is of great value and adds to Joan’s authority. As many believe, all language contains unconscious material; therefore, her method of retrieving that material is just an extreme form of the process that most authors share. [11] These intertexts and this mode of interpretation make me confident that Joan’s experience with automatic writing is relevant to our understanding of any kind of writing, even if her method is an extreme form of accepting the role of mediator.

As is the case with most creative writers, Joan is the first to frame her writing as literature. She is the one to designate the lines that she wrote in a trance and in full consciousness as poetry after she decides to quit automatic writing. Shocked by a terrifying experience in the trance, she terminates the experiments, “But I was left with the collection of papers” (Lady Oracle 224). Calling her work “papers” stresses that what she holds is a physical object that should not necessarily be seen as poems. Further social sanction, she suggests, is needed to transform the papers into literature. She is the one who needs to take the first step by “typ[ing] them up” as poems and sending them to a publishing house that accepts poetry. Here I am making an argument about the social nature of art similar to the one that helps to legitimize Marcel Duchamp’s found objects. It is Duchamp’s status as an artist that allows him to place a urinal in an art gallery and frame it as art, just as it is Joan and her publishers who have the power to frame text as poetry. In Latourian terms, we might say that the artist mediates between an object and consumers, creates a link between them that enables a new social structure around the object. The fact that a stricter Latourian than I would put more emphasis on non-human actors such as a page or urinal does not mean that my imaginary Latourian would disregard the role of the artist. The scenes in the novel in which editors, publishers, and journalists take partial control over the shape and reception of the poems highlight how seeing Joan as a node in a network is relevant even without the issue of automatic writing.

She seeks, distorts, interprets, and frames the text that comes from another source. Joan mediates between what she thinks of as the world of the dead and her living readers. It is clear in the novel in a way not explicit in Negotiating with the Dead that, even when performing what Hix calls a transcendental mode of authorship, a writer is much more than an intermediary implement for another’s will. Yet Joan’s work does not discredit how her mother (as ghost or projection) is also the author of the poem. In a study of the methods of attributing authorship, Harold Love defines the “precursory author” as “anyone whose function as a ‘source’ or ‘influence’ makes a substantial contribution to the shape and substance of the work” (40). Through this definition it is clear that if we accept a supernatural reading of this novel, then the mother is the coauthor of the poems, much like a public person is a coauthor of an autobiography whose final text is actually ghostwritten by a professional writer. Joan is the ghostwriter for her ghostly mother.

Joan shows that she is willing to take part in various arrangements of coauthorship and ghostwriting when she decides to use a pseudonym for her romance fiction. As in actual ghostwriting, the name that appears on the title page is not the author’s name. Joan does not pick an invented name but adopts the name of a real person, her aunt Louise K. Delacourt (Lady Oracle 157). Joan ghostwrites for Louise, who receives the authorial credit. Analogously, Joan is considered the author of the words that the ghost of her mother wrote. In this constellation, though, the aunt who is named on the cover is deceased, and it is the living Joan who does all of the actual writing.

In the much later novel The Blind Assassin (2000), the narrator-protagonist, Iris Chase, provides a dead author for her novel by similar means but under even more dramatic circumstances. Iris, married to a cruel and vindictive industrialist-cum-politician, writes a novel partially based upon her love affair with a communist. She cannot possibly publish this novel under her own name. Doing so would be tantamount to confessing her adultery publicly. Presumably, she could have invented a pen name, but she chooses a path with more symbolic resonance. After her sister, Laura, dies in a car crash (perhaps a suicide), Iris publishes the novel — also named The Blind Assassin — as a posthumous work written by Laura, who later becomes a tragic cult figure. Iris protects herself by attributing the novel to Laura. Like Joan, Iris becomes a ghostwriter, in her case by putting someone else’s name to her own text. Toward the end of the novel, the last two pages of her first-person narrative, Iris writes about her decision: “It was no great leap . . . naming Laura as the author. . . . I can’t say Laura didn’t write a word. Technically that’s accurate, but in another sense — what Laura would have called the spiritual sense — you could say she was my collaborator. The real author was neither one of us: a fist is more than the sum of its fingers” (512-13). Huebener reads this sentence as an indication that Atwood’s ruminations about authorship are applicable to the lives of non-writers, the overall argument of his essay (116). I think that the passage should be seen as a statement about how non-writers are part of the authoring process, writers helping to mediate their voices. The word spiritual sends us back to the possibility of communing with the spirits of the dead, but the word, especially in the context of The Blind Assassin, does not restrict us to ghosts. Rather, it opens up a world of different kinds of coauthoring, some more tangible than others. Atwood indicates again how writers, those who put pen to paper, mediate between different actors, some of whom might easily be considered authors too.

The issue of coauthoring and ghostwriting might have an additional turn of the screw in Lady Oracle because Joan, who has faked her own death, as we know from the first page, is not the only author of the narrative that we have as Lady Oracle. The narrative frame and the implied narratee might lead us to this further layer of source-mediator relations. Not all novels have clearly defined fictional narratees; most of them do not, so it is not necessary that Lady Oracle should have one. [12] However, the last chapter of the novel suggests that Joan does have an idea about who might be reading or listening to her story. [13] One way, perhaps the simplest, is to see the novel as a memoir that Joan is planning to publish when she returns to Canada. Her return becomes necessary because the friends who helped her to fake her death are suspected of her murder, and, as we learn in the final chapter, a reporter has found her in Italy. However, both Shuli Barzilai and Sherrill Grace present different and convincing ways to look at the narrative situation: the narrative is spoken with the aim of becoming a text written by another person, the reporter who discovers Joan in Italy (Barzilai, “Bluebeard” 267; Grace 192). Thus, the reporter, who followed Joan there, needs to be imagined as her messenger. She needs such a medium because she herself is deeply associated with death by faking her drowning and performing rituals of self-mourning. [14] The reporter, much like a spirit medium, will not only report what she says but also mediate and reshape it. Thus, in Lady Oracle, Joan is not only a medium and ghostwriter but also a ghost. The spirit medium who channels her voice to the realm of actual readers is Atwood herself of course. She is also a node in this system, though one with much power to act on it.

I want to go back to “Context Stinks,” in which Felski takes us a long way in reassessing literature as a continuous network in which the text is part of the world, transforms it, and is transformed by it. She gives us a much better model than the traditional dichotomy of text/context. Felski also highlights the limitations of a historicist reading that relegates the text to the period in which it was written, instead suggesting that the existence of the text after it was written or published be taken into account, especially its present relevance. Still, in making a powerful argument for the text’s post-publication network, she pays too little attention to its pre-publication itinerary, its assemblage or composition. Perhaps she reasoned that authorship, influences, and source studies have enough defenders. However, a full rethinking of literature as a network needs to take authorship and other pre-publication mediations into account.

Furthermore, read together with Latour, Atwood’s novel shows us that conceiving of the author as a medium, or in broader terms a node in a network, does not cancel out her role as an actor. This view lends new legitimacy to the study of authors, just as Felski’s view gives new legitimacy to the study of readers. Even elements put off limits such as biography or intention become legitimate again because they are part of the network that builds the persona of the author and therefore the assemblage of the text. At the same time, this argument shifts the focus of what we think of the kind of work that authors do. An author, like any other individual, is also a network of various physical and ideational components. The spirit medium is an image of the more complex process of bringing together material from a diverse network of agents: some within the author, some outside the author, all connected.

This spotlight on the author as one of many nodes of literary communication — of the writer as secretary, medium, editor — should also encourage more attention to other nodes in the network of composition. Such work is already being done, of course. I have cited York’s crucial contribution. We must not stop, though, with human agents. Instead, we should study things, objects, as mediators as well. The attention given in Lady Oracle to the things used in the automatic writing process (paper, pencil, candle, mirror, dictionary) suggests that these elements merit further consideration. With the increased interest of literary critics in Latour, as well as the rapidly mutating technologies of writing and publishing, these kinds of questions will be asked more often by critics and creative writers alike. The answers are likely to give us a much better understanding of how authors and authorship function among different mediators all taking key roles in the production of texts, none cancelling out other mediations.