Linguistic Code vs. Bibliographical Code

In the AI Story

McGann developed the distinction during decades of editorial work on Romantic and Victorian poetry, where the material differences between manuscript, first edition, and subsequent editions often carried semantic weight that conventional textual criticism had ignored. A poem's lineation, the weight of its paper, the typeface chosen by a compositor — these were not incidental features awaiting correction to match the author's intention but constitutive elements of the text's meaning. Emily Dickinson's dashes, normalized into periods by well-meaning editors, altered the rhythm and therefore the meaning of her poems. The alteration was not a correction of error but a substantive change that the conventional focus on linguistic content had rendered invisible.

The framework's deepest claim is that the two codes cannot be hierarchically ordered. The linguistic code is not primary and the bibliographical code supplementary; they operate together in producing the reader's experience of the text. A reader who picks up a clothbound university-press volume processes information about authority, scholarly rigor, and intended audience before reading a word. That processing shapes how the subsequent words are received. The bibliographical code is doing semantic work that the linguistic code does not and cannot duplicate.

Applied to AI text, the framework reveals that machine-generated prose has its own bibliographical code: tonal consistency across long passages, structural regularity, syntactic fluency that does not vary with the difficulty of the material, and the absence of what might be called textual grain — the productive awkwardness that marks a mind reaching for expression at the edge of its capacity. These features are not defects. They are information. They communicate something about the text's conditions of production that readers absorb as part of the reading experience, whether or not they can articulate what they are absorbing.

The stakes are practical as well as theoretical. When the bibliographical code of polished prose no longer reliably signals careful thinking — because polished prose can now be produced at near-zero cost — readers must develop new interpretive skills. They must learn to read form as information, to notice when the code is communicating something that the content does not support, and to resist the automatic attribution of authority that material polish has always commanded. These are the skills McGann's scholarship has been developing for forty years, now required not of specialists but of any reader navigating the post-exposure textual environment.

Origin

The distinction crystallized in The Textual Condition (1991), but its roots extend through McGann's earlier editorial work on Byron and his archival research for the Rossetti Archive. The encounter with texts whose meaning changed decisively across different material instantiations forced a theoretical vocabulary that the dominant editorial tradition did not provide. Greg-Bowers copy-text theory treated material variation as noise to be filtered toward the author's final intention. McGann's framework treated material variation as signal, carrying meaning that the intentionalist framework systematically erased.

Key Ideas

Two simultaneous codes. Meaning is produced by the interaction of linguistic and bibliographical dimensions, not by the linguistic alone.

Materiality as semantic. Typeface, binding, layout, and cover design carry meaning independent of the words and shape how the words are received.

No hierarchy. The bibliographical code is not decorative or subordinate; it operates in the reader's experience before conscious evaluation of content begins.

AI's new code. Machine-generated prose constitutes a novel bibliographical signature — smoothness, consistency, the absence of struggle — that communicates about origin regardless of content.

Reading form as information. The practical skill the framework demands is attention to material presentation as a text requiring interpretation alongside the linguistic text.

Debates & Critiques

The framework has been contested by textual scholars who preferred the cleaner analytical territory of intentionalist editing, where the goal is recovering the author's final wishes and the bibliographical accidents can be set aside. McGann's response has been consistent: the accidents are not accidents, and the author's final wishes, if they exist at all, are always shaped by the material conditions of production. The debate is not about evidence but about what counts as the text — and McGann's framework insists that the material form is part of what counts.