Editorial Invisibility

In the AI Story

The convention began to strain in cases where editorial intervention crossed from refinement to co-creation. The Lish-Carver case—where editor Gordon Lish cut Raymond Carver's stories by fifty percent or more, rewrote endings, and altered emotional registers—exposed that the convention depends on invisible boundaries. When editing becomes extensive enough that the published work's aesthetic character reflects the editor's sensibility as much as the author's, the question of authorship becomes genuinely difficult. The literary world never fully resolved this question, settling instead into uncomfortable acknowledgment that some collaborations exceed the convention's capacity to manage them.

The convention's ethical coherence depends on temporal structure: editing operates on completed drafts, meaning the editor's contributions arrive as discrete interventions the author can accept or reject. This discreteness allows the collaboration to be tracked in principle even if it remains invisible in practice. The author knows what the editor suggested and what the author chose. The invisibility conceals specific decisions but not the fact that decisions were made. AI collaboration eliminates this discreteness: the machine's contributions arrive concurrently with the author's thinking, integrated into the ideation process rather than applied to finished text. The author who tries to identify which ideas were influenced by Claude faces not a problem of record-keeping but of cognitive archaeology—excavating boundaries that were never clearly drawn.

Lesser's editorial practice at The Threepenny Review represents the tradition at its best: substantial contributions to manuscripts, genuine shaping of final texts, complete invisibility to readers. The convention survives in her practice because the contributions remain responsive. Lesser reads what the author has written, forms a response based on her encounter with the text, and suggests changes that serve the author's vision. The suggestions are discrete, identifiable, and optional. The author retains authority. The invisibility remains ethically stable because everyone understands what it conceals: not co-authorship but skilled assistance.

AI collaboration produces what this volume calls ontological invisibility: the collaboration is invisible not by convention but by nature. The contributions cannot be made visible because they were never separate from the author's thinking. The text that emerges carries the marks of both human and machine without allowing retrospective separation. The conventional acknowledgment—"written with Claude"—is honest but inadequate: it names the collaboration without describing it, because the collaboration cannot be described in the vocabulary the convention provides. The author who says "my editor improved this chapter" is making a claim about a discrete contribution. The author who says "I wrote this with Claude" is acknowledging a partnership whose products cannot be cleanly attributed.

Origin

The convention of editorial invisibility emerged with modern publishing in the late nineteenth and early twentieth centuries as the editor's role professionalized. Earlier literary production—patronage systems, author-publisher collaborations, collective authorship—had different attribution norms. The modern convention arose alongside Romantic authorship: the ideology that texts originate in individual creative consciousness. The editor became invisible precisely because the ideology required a visible author—a single originating mind—and collaborative production threatened that picture.

The convention persisted for a century because it served everyone's interests: authors received sole public credit, editors received professional recognition within the industry, publishers benefited from the author-as-brand model that individual attribution enabled. Occasional cases—acknowledged co-authorships, ghostwriting scandals—tested the convention without overturning it. The Lish-Carver revelation in the 1990s produced the most sustained questioning, but even that case settled into acknowledgment of collaboration without formal revision of attribution practices.

Key Ideas

Conventional versus ontological invisibility. Traditional editing hid contributions by professional norms (conventional); AI collaboration hides contributions by the nature of concurrent generation (ontological).

Discreteness assumption. The convention assumed editorial contributions were identifiable acts—cuts, additions, structural suggestions—that could be tracked even if they remained publicly invisible; AI dissolves this assumption.

Responsive versus initiatory. The editor's secondary role—responding to the author's text—justified invisibility; when collaboration becomes initiatory (generating ideas, not just refining them), the ethical foundation weakens.

Temporal structure collapse. Editing on completed drafts created temporal separation allowing attribution; editing inside the thinking process eliminates the separation and makes attribution impossible.

Acknowledgment's new inadequacy. The conventional gesture—naming the editor in acknowledgments—remains the right practice but points to a collaboration the convention's vocabulary cannot adequately describe.