Internet has shaped a new relationship with knowledge that has given rise to new forms of encyclopaedism. These are linked to the idea that we could access an exhaustive knowledge of reality thanks to a real-time archive of all the texts and images we produce. The computerisation of knowledge immerses us in an incessant digital flow that gives the impression of being able to ‘archive everything live, (...) in the present tense’ (Bertrand Gervais), with no delay or lag between the event and its preservation. Indeed, archiving in the present not only means preserving traces and remains, but also instantly recording all the texts and images we produce on a daily basis. Obsessed by the desire to capture time in its immediate dimension, digital culture is motivated less by the preservation of traces of the past than by the endless accumulation of unstructured, insubstantial data.

This development is not without consequences for our relationship with time and history, and for our ways of telling and reading. The transition from ‘digital reason’ to ‘graphic reason’ (Jack Goody) has generated new ways of presenting and organising knowledge, affecting both the scholarly forms of encyclopaedism (atlas, dictionarie, inventories, encyclopaedic novels, etc.) and their literary appropriations. Whether in digital or paper format, contemporary forms of encyclopaedism oscillate between a desire for exhaustiveness and an awareness of its impossibility, between readability and unreadability, mimicry and resistance to the data regime. This ambivalence in turn generates new ways of writing and reading knowledge, which we will try to highlight, on the basis on the work of Judith Schalansky, who, from one book to the next, has explored several forms of totalising knowledge: encyclopaedias, archives, atlases, schoolbooks, and so on. We will show how she reinvents these forms to exploit their cognitive and architectural potential, but also to thwart their principles and effects.

In his Poetry in a Global Age (2020), Jahan Ramazani presents Matt Rasmussen's poem "Reverse Suicide", which is composed like a film played backwards. Which the reader will probably realize sooner or later also because of the title. Until that moment the poem seems incomprehensible and strange, after this "media shift" it suddenly becomes clear and understandable. The second moment of recognition is that such a poem could only have been written after the advent of film, which brings a certain type of media capture of movement (and much of it can be seen just in the reverse movement of film). In my paper I will try to show some manifestations of this inner mediality of poetry, which is both virtual (poetry prefigures possibilities of other media) and influenced by the real effects of other media.

In its origins, as an oral expression or technique, poetry was a significant and dominant medium (in a narrower sense of a channel), especially of collective memory, as the research of M. Parry, Eric Havelock and others have shown for Greek and Western poetry. Poetry, however, gradually lost this role with the advent of writing and the development of writing techniques, becoming a medium in the broader sense of cultural practice (on the distinction see Baetens). The subject of my interest is poetry and its techniques as a secondary channel that mediates other media. In the first part of the paper I will briefly discuss the ancient ekphrasis, which is an example of complex virtual visuality. In Homer, Hesiod, and later authors we find verbal descriptions of representations (depictions on a shield, painting, etc.) that are characterized by great dynamism and use verbal visualization to capture scenes that would be impossible to capture in a static representation (such as a painting or a photograph) and, in some cases, even in a theatrical scene unfolding in time. In some ways, this verbal representation of visual imagery virtually contains the possibilities of, for example, film technique (not only movement, but also the trick manipulation etc.). In the second part of this paper, I will discuss some examples from modern poetry that work in contrast with the real influence of other media (such as Rasmussen's poem). Whereas in the first case poetry expanded the field of visuality, in this second form other media expand the field of poetry.

References:

Baetens, Jan. 2025. "I.2.Mediality and Materiality of Lyric." In Poetry in Notions. The Online Critical Compendium of Lyric Poetry, edited by Gustavo Guerrero, Ralph Müller, Antonio Rodriguez and Kirsten Stirling. https://doi.org/10.51363/pin.728c

Antonio Rodriguez and Kirsten Stirling. Lyre multimédia. Études de lettres, 2022, no 319. https://journals.openedition.org/edl/3969

The term ‘distant reading’ came up at the beginning of the new millennium and is generally attributed to Franco Moretti (cf. Moretti 2000). As a counter-concept to ‘close reading’ it applies computational methods to analyse large amounts of literary data. Instead of the detailed reading of individual texts, which focuses on the writing and reading subject, ‘distant reading’ processes masses of text according to certain recurring patterns or predetermined criteria. As such a tool that can work on much more texts than a single person would ever be able to, in terms of time, ‘distant reading’ takes on a new dimension in the current debate on AI-generated texts. Apart from the reading function of LMMs proposed by Hannes Bajohr (cf. Bajohr 2024), it could be a ray of hope in the “textpocalypse” (Kirschenbaum 2023) conjured up by Matthew Kirschenbaum. With ‘distant reading’, the flood of AI-generated texts that, according to Kirschenbaum, will soon be invading the Internet and our lives, would be manageable to a certain extent. We could at least filter recurring words, phrases or topics from the masses of text and thus get an idea of what they are about. However, that’s where the cat bites its tail, as the AI-generated texts were produced according to the same principle: on the basis of the most frequently used word sequences that result from the machine's sifting through huge text corpora. The machine “reads” what the machine has “written” in order to “write” new text from it.

My paper rtakes up this point and outlines an active role humans can have in this seemingly endless nonhuman feedback loop. Instead of being paralyzed by the oncoming “textpocalypse”, ‘distant reading’ is to be developed as a poetic instrument that, beyond identification-driven individual readings, enables a productive approach to unmanageable masses of text. I will show what this writing with and for ‘distant reading’ can look like using the collection "Halbzeug" (2018) by Hannes Bajohr and contrasting it to Raymond Queneau's "Cent Mille Milliard de Poèmes" (1961). Bajohr filters specific text corpora with ‘distant reading’ and collages new texts from the result, whereas Queneau develops a text, that can literally only be read at a distance or by a machine – because human lifetime is simply not long enough to do so.

- Bajohr, Hannes (2024): “Große Sprachmodelle. Machine Learning als Lese- und Schreibermöglichung“, in: Zeitschrift für Medienwissenschaft. 16:31, pp. 142–146.

http://doi.org/10.25969/mediarep/23149.

- Kirschenbaum, Matthew (2023); “Prepare for the Textpocalypse. Our relationship to writing is about to change forever; it may not end well”, in: The Atlantic. 08.03.2023

https://www.theatlantic.com/technology/archive/2023/03/ai-chatgpt-writing-language-models/673318/

- Moretti, Franco (2000): “Conjectures on world literature”, in: New Left Review I:238.

https://newleftreview.org/issues/ii1/articles/franco-moretti-conjectures-on-world-literature