Galit Wellner’s (2017) concept of writing relations provides a useful starting point for examining how generative AI transforms not only who writes, but how writing unfolds in time. Wellner develops writing relations to distinguish inscription from the interpretive activity privileged by postphenomenological hermeneutic relations. Writing, for Wellner, is productive: through the mediation of technology, it creates a textual trace that enters and alters the world. This leads her to expand the conventional human–technology–world relation into the quad-scheme I—tech—text—world, allowing the technological artifact, the produced text, and its relation to the world to be considered separately.
This framework takes on particular significance when Wellner turns to algorithmic writing. In her third variation, (I—) tech → text → world, the human is bracketed while technological intentionality proceeds toward the production of text. As Wellner (2017) explains, “Bracketing the ‘I’ means the human withdraws to the background at the stage of writing” (p. 223). Although Wellner formulates this withdrawal primarily in terms of machine writing and technological intentionality, it also raises a temporal question: what happens to the experience of writing when the human withdraws precisely during the period in which the text comes into being?
This essay takes that question as its point of departure. I argue that generative AI produces a temporal flattening of writing by compressing formerly differentiated moments of ideation, formulation, inscription, anticipation, and revision into an encounter with text that appears already composed. The consequence is not simply acceleration. As the writer becomes detached from the temporal unfolding of composition, the architecture of writing itself is reconfigured. This temporal flattening ultimately gives rise to technological post-temporal time-consciousness: a condition in which writing appears not as something composed progressively through lived time, but as something surfaced from an algorithmic field of linguistic probabilities. The bracketed writer is therefore displaced not only from the immediate agency of inscription but from the temporal flow through which authorship has traditionally been experienced.
Temporal Flattening: From Authorial Withdrawal to Temporal Withdrawal
Wellner’s bracketing of the “I” becomes especially significant when considered temporally. She describes writing as an activity occurring within a particular duration, that is a text may be written at once or over weeks, months, or years, and elsewhere situates writing, displaying, and reading as stages “stretching along a timeline” (p. 211). Writing, then, is not merely the production of a textual trace but a process through which that trace gradually comes into being. Yet in Wellner’s (I—) tech → text → world, the human withdraws at precisely this stage of inscription while technology assumes the immediate work of producing the text. With generative AI, this withdrawal becomes particularly pronounced because several compositional operations can be delegated simultaneously.
The resulting transformation can be called temporal flattening: the compression of temporally differentiated acts of ideation, formulation, inscription, revision, and anticipation into an encounter with text that appears already composed. Conventionally, the writer inhabits the text’s becoming, moving recursively between what has just been written, what is being written, and what might follow. Generative AI interrupts this duration. A prompt can produce in seconds an organized passage whose words, transitions, and argumentative trajectory the writer did not progressively traverse. The writer thus encounters the sentence without necessarily experiencing the search for its words, or the argument without anticipating each movement through which it develops. Temporal flattening is therefore not simply acceleration. Acceleration would mean that the same compositional sequence occurs more quickly; generative AI can instead remove portions of that sequence from the writer’s lived experience. Computational processing still unfolds chronologically, but it does not constitute the phenomenological duration of human composition. What becomes flattened is consequently not chronological time itself, but the writer’s experience of its differentiated compositional moments.
Wellner’s bracketing of authorship can thus also be understood as a bracketing of duration. The human remains temporally active before and after generation: prompting, reading, evaluating, selecting, or revising, but withdraws from the interval in which tech → text occurs. Writing thereby shifts from continuously inhabiting the emergence of text toward alternating between human interventions and algorithmic textual arrivals. The question raised by temporal flattening is therefore ultimately phenomenological: what happens to the writer’s experience of past, present, and future when the text is increasingly encountered not as something unfolding through one’s own temporal activity, but as something that has already arrived?
From Temporal Flattening to Post-Temporal Time-Consciousness
Temporal flattening becomes philosophically consequential when it alters not only the organization of writing but the writer’s time-consciousness. Temporal (time) consciousness is our lived awareness of temporality, the way consciousness experiences itself and its objects as flowing through past, present, and future (Thomas, 2023). If generative AI occupies part of the interval in which language is formulated and anticipated, the writer encounters text differently from one who experiences its gradual emergence through composition. This disruption of compositional continuity opens toward what I call technological post-temporal time-consciousness.
Husserl’s account of internal time-consciousness (Jacob, 2019) provides a framework for understanding this shift. Experience unfolds through the interrelation of retention, primal impression, and protention: retention holds the just-past within consciousness, primal impression constitutes the living present, and protention anticipates what is about to occur. Writing enacts this structure particularly clearly. What has just been written remains present while the writer composes the current sentence and anticipates what might follow. The text therefore emerges through a continuous interplay of the just-written, presently written, and not-yet-written. Generative AI disrupts this continuity by intervening between what has been written and what has yet to be written. A generated passage may possess linguistic and argumentative continuity even though the writer has not lived through the sequence of inscriptions and anticipations that produced it. Linguistic continuity is thus preserved in the output while phenomenological continuity is interrupted in composition.
This interruption first transforms retention. In human writing, the just-written remains present as part of the writer’s own lived past. With generative AI, however, the linguistic past conditioning what appears next exceeds the writer’s individual experience. Generation draws upon statistically sedimented linguistic patterns rather than a phenomenologically remembered past. Retention consequently shifts toward what might be called technological sedimentation: a de-individualized textual past that conditions the generated present without itself belonging to an experiencing “I.” Wellner’s bracketed author therefore acquires a temporal dimension. In (I—) tech → text → world, the human withdraws at the stage of writing. Under GenAI, that withdrawal extends into the temporal constitution of the text. The writer is displaced not only from inscription but partially from its past, encountering writing conditioned by a textual history that exceeds their own lived temporality. Temporal flattening thus begins to develop into post-temporal time-consciousness: the text remains sequential, but its past need not be the writer’s lived past, its present need not coincide with the writer’s act of inscription, and its future, as the transformation of protention reveals, need no longer originate primarily in the writer’s own anticipation.
Primal Impression, Protention, and Chronological Exile
If retention shifts from the writer’s lived past toward technological sedimentation, primal impression, the living present, undergoes a related displacement. In conventional composition, the emergence of writing and the experience of composing it largely coincide: anticipated words become inscriptions within the writer’s present activity. With GenAI, these moments separate. The writer initiates generation and encounters its result, while the linguistic operations connecting the two occur computationally. The living present of writing therefore becomes distributed across prompting, algorithmic generation, textual appearance, and human reception. In Wellner’s terms, the bracketed “I” thus withdraws not only from inscription but from part of its present: the generated sentence arrives as a present experience whose formulation has already occurred elsewhere. This produces an inversion of intentionality. Rather than human intentionality progressively giving rise to the text, generated text becomes an object toward which intentionality is subsequently directed: the writer reads, evaluates, rejects, revises, or regenerates what appears. The writer consequently occupies the present less exclusively as an inscriber and as an interpreter of textual arrivals.
The more consequential transformation, however, concerns protention, the anticipatory orientation through which the present opens toward what comes next. Writing ordinarily depends upon this not-yet-written horizon: words anticipate possible continuations, sentences project toward subsequent sentences, and arguments develop through possibilities the writer has yet to realize. Generative AI intervenes directly in this horizon by materializing possible continuations before the writer has consciously formulated them. The human reacts to the generated text rather than phenomenologically anticipating and inscribing what should follow. Protention is therefore not merely accelerated but partially outsourced: the machine populates the future of composition with textual possibilities that the writer subsequently evaluates. Here temporal flattening reaches its most radical form. What belonged to the not-yet-written can return immediately as generated text, contracting the distance between anticipating text and encountering it. The writer moves from What will I write next? toward Is this what should come next? Possible textual futures become present objects of selection rather than possibilities reached solely through the writer’s own anticipatory activity.
This displacement produces chronological exile. The writer does not literally leave chronological time but becomes estranged from the temporal flow traditionally constitutive of composition: retention is supplemented by technological sedimentation, the living present is distributed across human and computational activity, and protention can be pre-empted by algorithmic next-token prediction. The writer occupies the edges of composition, initiating before generation and interpreting afterward, while technology occupies part of its temporal middle. It is through this reconfiguration that temporal flattening develops into technological post-temporal time-consciousness. Post-temporal does not imply the disappearance of chronological time; rather, past, present, and future are no longer organized exclusively through the writer’s lived temporal consciousness. Writing consequently acquires an oracular temporality: instead of progressively composing text through time, the writer consults an algorithmically structured field of linguistic possibilities from which possible textual futures are surfaced into the present. The movement initiated by Wellner’s bracketed author therefore culminates in a temporal displacement where writing shifts from an act of becoming toward an encounter with prefigured linguistic possibility.
References
Jacob, S. (2019). Husserl’s model of time-consciousness, and the phenomenology of rhythm. In P. Cheyne, A. Hamilton, & M. Paddison (Eds.), The philosophy of rhythm: Aesthetics, music, poetics (pp. 291-306). Oxford University Press.
Thomas, V. C. (2023). The essentials of Husserl: Studies in transcendental phenomenology. Vernon Press.
Wellner, G. (2017). I-media-world: The algorithmic shift from hermeneutic relations to writing relations. In Y. van den Eede, S. O’Neal Irwin, & G. P. Wellner (Eds.), Postphenomenology and media: Essays on human–media–world relations (pp. 207–227). Lexington Books.
