Poems entering the public sphere from its margins

See more about the poesiefestival berlin on their website.

The awarding of the most important prize for contemporary German-language poetry, the Peter Huchel Prize, is normally announced without the participation of the wider public. The annual news round-up from the public broadcaster Südwestrundfunk, which sponsors the prize together with the state of Baden-Württemberg, receives only a few likes and hardly any comments on social media. The discussion of the decision takes place solely in poetry circles as well as the feuilletons of major newspapers and culture sections of broadcasters’ websites. In 2023, however, it was different. The post about the jury’s decision in favour of Judith Zander and her book im ländchen sommer im winter zur see generated a public reaction without a precedent in contemporary poetry in German-speaking countries, oscillating between a debate and a shitstorm: the post received around 800 reactions – half laugh smileys, half likes and hearts – and well over 1700 comments (as of 1 May 2023).

On the one hand, this was a typical social media phenomenon that reinforced itself once it got rolling. But how did it come about in the first place? After a spot check, Frank Milautzcki assumed that there might have been a call to hijack the page among conspiracists and right-wing populists. This cannot be ruled out, but the poem has been criticised in left-wing and liberal circles, too. Others suspected that it may have to do with the gender of the laureate. But this would be implausible as the sole reason, as there have been six female laureates since 2010, without any such reaction by the wider public. This year’s post differed from those of the previous years in that instead of a statement or a shorter quotation, an entire poem of 25 lines was printed. SWR did not bother to post a poem from the award-winning volume, but copied an older one from lyrikline.org: grundlegende.

The reactions to this post were mostly negative and mainly related to the poem’s incomprehensibility: users did not understand why everything was written in lowercase and set without punctuation marks, why foreign vocabulary was used, what individual words, their combination, single lines or the poem as a whole were supposed to mean. Confessions of not being able to understand even after multiple readings were accompanied by repeated references to well-known parodies of unintelligible contemporary poetry and music by Loriot and Hape Kerkeling. In addition, there was a plethora of comments mocking and defaming the text as nonsense as well as contemporary poetry as elitist and academic, pasting own poems as parodies or alternatives, and in some cases viciously insulting the author, whom the commentators had not previously known.

On the other hand, voices were raised, mostly by fellow poets, who defended the poet and her poem as well as contemporary poetry in general with interpretations and explanations. The defenders, too, argued ad hominem and accused the spiteful commentators in this and other threads of stupidity, ignorance, lack of respect, resentment, anti-intellectualism, herd instinct, right-wing populist sentiment or misogyny. Soon afterwards, Judith Zander herself spoke out on Deutschlandfunk and expressed suspicions of anti-lyric reflexes and envy among the commentators.

Several contributions by poets and poetry critics in online magazines and literary supplements of the print ones reacted to these Facebook disputes. However, the question of why this debate came about and what can be learnt from it remains open in German-language poetry discourse. In the following pages, I would like to offer some reflections on the conditions under which this clash of the poetry scene and members of the public on social media took place, as well as some reasonable conclusions.

I think it would be instructive to pursue the idea that this debate was not so much about the poet and her poem as it was about the form of much of contemporary poetry in general. The public response reveals a dissonance between the expectations and what the poem delivers. It calls for reading in three ways: 1) as a small literary form that fits the dispositif of short content on social media, as it is evident in phenomena like Instapoetry, 2) as a text singled out by a public broadcaster as an item of general interest, and 3) as a Facebook post whose quick and easy reception is the algorithmic normal of this ad-financed platform.

At the same time, though, ‘grundlegende’ resists the frictionless application of a basic competence that is part of basic literacy. As a poem rich in line jumps, it does not read fluently a) like most Instapoetry or a conventionally narrated page-turner, b) a news article, c) a typical piece of content adapted to the minimal attention span of social media. The poem’s failure to meet the recipients’ expectations triggered aggression. Hence, the public reactions to Judith Zander’s poem show which cultural expectation of poetry continues to be effective: that poems are in principle for everyone, indeed that they are the form to mediate ‘between me and my people’, as Ben Lerner observes in his The Hatred of Poetry from 2016 (p. 16).

All the subtlety of language that one can experience in the poem, if one takes one’s time and perceives accurately, was obscured in the dispute by the paradigm of appropriation through quick understanding, which the social media dispositif insists on. Much got almost entirely lost, for example, the highly differentiated sound space of the poem with internal assonances, the complex play of various layers of meaning, which begins with the title that can be read in three ways (as ‘underlying’, ‘basic legend’, perhaps also as ‘the end of laying ground’); the liaisons of German with other languages, the surprising and insightful connections of technical and everyday vocabulary, the interplay of presence and absence, the brightness and colour values of words, their hardness and softness, intertextual references to Kafka and others, and much more.

Therefore, the Facebook thread provides evidence for the continuous crisis of poetry-critical education – many people do not know how to approach a contemporary poem openly and critically. Several users regarded themselves as insufficiently capable of dealing with the poem and explicitly asked for help and advice. Other judgmental comments can be understood as an attack on the poetic itself: opinions aggressively fired at a textual form that refuses to be about expressing opinions.

What conclusions are to be drawn from this? Logically, there are three possibilities for poets and the larger public: 1) poets should open up to a wider audience; 2) the public should be more accommodating to contemporary poetry; 3) the two should remain separate.

All three positions are held within the German-language poetry scene including the spoken word and slam scene: on the one hand, there are complaints that poetry is too elitist, academic, self-referential, incomprehensible, or too German, bourgeois and homogeneous. On the other hand, poets complain that poetry is notoriously underrepresented and neglected in the public sphere, which is why the masses, like Facebook commentators, judge what they do not understand instead of acquiring knowledge of and openness to poetry first. The third position, ironically defended some years ago by Hans Magnus Enzensberger, is taken even further today, for example when Sabine Scho, in the face of the aggressive Zander debate, pleads for safe spaces for poetry against the ignorance of the crowd.

The latter position is supported by the fact that the relative freedom of poetry from adapting to audience expectations, which is connected with its economic irrelevance, apparently flourishes in marginal biotopes such as the self-referential poetry scene and has produced an astonishing aesthetic variety. Would the boom in contemporary poetry in the German-speaking world in the first two decades of the 21st century, which the literary scholar Christian Metz celebrated in his major 2018 study Poetisch denken (Thinking Poetically), have been possible if poets had been concerned with reaching a broad audience and securing a livelihood? But the decoupling of poetry from its broad public reception also implies the continuation of its relative irrelevance for important social and political debates. I want to argue that positions 1) and 2) should be combined.

What would it mean for poetry to accommodate the general public? For one thing, the poetry scene would be as diverse as the society and would lower its access barriers. In Germany, it in fact has become more diverse recently. Most influential German-language poets are female; post-migrant, Jewish and queer voices are far more present than they were at the beginning of this century, and not all poets are academically trained. Yet, these politically important changes do not necessarily lead to more comprehensible poetry. How can exclusions be countered here without giving away the aesthetic diversity and complexity that contemporary poetry has built up since the 1990s?

One way to make the qualities, practices and forms of the poetic generally more accessible and increase openness to poetry is to structurally expand the possibilities of experiencing poetry. In German public schools aesthetic practice is hardly ever taught in literature lessons (unlike music and art lessons). Poetry is taught almost exclusively as an object of interpretation with standards for marking tests (many commentators in the Huchel prize debate therefore referred to their school experience with Goethe and so on and pitied students who may eventually have to interpret Zander's poem). Programmes for poetic education have been set up recently, mainly at the Haus für Poesie (House for Poetry) in Berlin, where poetry workshops for children, teenagers and young adults are held regularly (also as part of the annual poesiefestival berlin). Other institutions such as the Lyrik Kabinett in Munich, the Friedrich-von-Bödecker-Kreis or the Literaturhaus Stuttgart also offer poetry workshops or training for teachers. Poedu is a project for young children in primary schools, which was first set up during the pandemic. Lyrix, the national youth poetry competition, awards monthly prizes for contributors between the ages of 10 and 20 and organises writing workshops at museums and schools.

Aesthetic writing practices, if they are included in the curricula for (primary) schools at all, are usually quite schematic and follow outdated poetics. Contemporary poets, coming from outside of schools, can encourage a more experimental, freer writing with reflection on poetic speech itself. These and other initiatives, which are (partly) collected on the lyriklab website launched at the end of 2022, are perhaps the most promising movement in poetry education that Germany has seen so far.

But poetic education has nowhere near the resources to reach many, let alone all, children and young people, especially those from educationally disadvantaged backgrounds. Or even adults of all ages. Large investments from the public sector are needed here, perhaps via a poetry fund, which the Netzwerk Lyrik, founded in 2017, advocates. Such investments would benefit the poetry scene and give poets more opportunity to engage with a wider public in order to close the gaps of poetic-critical skills. They would also benefit the democratic society as a whole, especially in this historical moment, as the current Facebook controversy over an allegedly incomprehensible poem is taking place at a tipping point in the development of artificial intelligence. The momentousness of this technology makes it all the more important to pursue poetic education on a much larger scale.

Meta, which owns Facebook, is one of the few corporate players that can still afford to train LLMs (Large Language Models), after a long time when this was mainly done by academic initiatives that now struggle to keep up due to the extreme increase in costs. LLMs like ChatGPT learn from large amounts of natural language of the entire networked world, from which they in turn generate language according to statistical probability. Machine learning tends to solidify language standardisation in a globalised stilus mediocris, in which conspicuous exceptions and deviations are increasingly eradicated. The understanding of digitally generated character strings is already amazingly trouble-free: texts are comprehensible, grammatically correct and stylistically appropriate.

What happens when text production from interaction with bots becomes the standard in the digital space? In fact, humans themselves are kind of generative pre-trained transformers (GPTs) when they speak, as they acquire language from many sources in their lives. But we can veer off the training track and open up to contingency, incoherence, perplexity, surprise and, hence, novelty. Now, we have started to trust super enhanced GPTs speaking for us in a fairly intelligible manner from a vast corpus of sources even if they produce bullshit. The amount of text stochastically generated from texts of the past is growing so rapidly as if to outpace the ecologically disastrous growth curves of the Great Acceleration: ChatGPT had 1 million users after 5 days, 100 million after 2 months, over 1.6 billion after 4 months – the fastest growth of a user base in tech history.

The fact that the technical production of texts is concealed in a globalised hegemony of style together with the economic motives of the producers can bring about a mystification of human intelligence. In the global proliferation of information, ways of speaking inherited from the problematic past, and the attached partly concealed ideologies with their implied racist, sexist, etc. value judgments, are taken over and making more room for biases and projections. Conversely, the general focus on automatically generated comprehensibility in digitised societies could lead to a more unconscious and uncritical approach to the possibilities, nuances and limits of textual language.

The creative as well as critical autonomy of each individual in their relation to language is facing the threat of further erosion if LLMs replace human authorship instead of being used critically as tools by human authors, just as representatives of digital poetry since the 1950s have used programs, algorithms and AI to produce fascinating contributions that challenge everyday wording and its apparent intelligibility. The nuances of human language usage face the threat of becoming even more unconscious. Manipulation and misdirection could become far more effective than they were when they relied on the human-driven social media algorithms which also provide statistical reproductions of the past by showing you content you’re likely to click on according to your history, denying users perspectives they would not immediately understand and would hesitate to spread further.

If the standard set by machine learning companies shapes the political sphere and parcels it out, according to ways of speaking and understanding, to types of opinion and evaluative attitudes, that would be ‘a democratic disaster’, as the poet, philosopher and literary scholar Hannes Bajohr recently pointed out. I agree with him that attempts to make political discourse more harmonious through the use of LLMs which ‘improve participants’ perception of feeling understood’ in a language ‘ironed out and set to the most middling version of itself’, seem rather dystopian. Ironically, the communication-orientated discourse of open societies is endangered precisely by the interference-free understanding of machine-generated meaning that smooths down critical and controversial discourse.

It’s no surprise that if one asks ChatGPT for a difficult-to-understand, fractured contemporary poem, the result is a rhymed, easily digestible conventional piece close to kitsch (I tried it several times in German and English). On the contrary, putting words in proximity in a phrase like ‘lichtkeimer’ (seeds that require light for germination), ‘an und für sich’ (the Hegelian ‘in itself and for itself’), ‘follower’, ‘acte de volonté’ (French: ‘act of will’) and ‘verlaufene’ (‘the dispersed’ / ’the ones who got lost’) as in Zander’s poem, is statistically improbable. Theoretically, it could be generated by a machine but not by current LLMs as the output is not normal because it is not average (and, hence, annoying to many users and ideologically useless to big tech). Contemporary poems can create a challenge for our understanding and a need to refine our hermeneutic competence and our (aesthetic) faculty of judgement which may include a readiness to link what is separated and separate what is linked.

However, we do not only lose a great potential for expression and critical distance to the increasing habituation to frictionless understanding of texts in standard styles, but also the ability to infer intuitively what expressive and cognitive potential we might be missing, what may shift the whole paradigm or allow us to think outside of all boxes already known to us. Yet, the current AI solves problems we know – this is a new problem. With a weakened sense of possibility (‘Möglichkeitssinn’, described by Robert Musil in his The Man Without Qualities), we become less motivated to uncover the echoes of histories of violence in languages and to discover treasures of illuminating spaces of feeling and thinking in languages and between them.

Searching for other ways of speaking and thinking, for a new sense of pleasure in what we can hear, see, feel and think in so many different ways in the medium of language that connects humanity, requires a practice of deviating from hardened linguistic conventions and its hegemonial tendencies. Only a slowed down, irritating, rhythmically interrupted process of understanding can deepen and expand comprehension.

No text form gives as much cause for this as poetry and its verses, line breaks, caesurae and ruptures with which it explores ways of speaking, feeling and thinking beyond the familiar. Playfully making and receiving somewhat complex poetry trains us to be ready to take detours around the paradigm of unambiguous opinion, the readily understandable and the repeatedly said, and resist the powerful deployment of language standardisation.

This can be irritating, absurd and hilarious – all of which is welcome in the spaces of poetry, to which AI as a tool can greatly contribute. Through poetry we can experience an aesthetic liberation from the power of language in language. When this liberation, with its possibilities, is realised, it becomes more conscious to poetised people. Therefore, poetic education for all, understood as an exercise in poetic speaking, writing, listening and reading and critical reflection, is a political task. Europeans should tackle it together.