Of Minds and Machines

  • Hannes Bajohr
  • ·
  • 2016-07-06

[This essay first appeared in Logbuch Suhrkamp.]

A year ago, Swantje Lichtenstein and Tom Lingnau asked a number of artists a simple question: Is the artist necessary for making art today? What became a slim volume of answers began with their project Covertext, which investigates literary conceptualism and appropriation. It was in this context that they first questioned what connection remains between artist and artwork – for if conceptual literature takes the idea to be more important than its execution, and if the pure appropriation and copying of other people’s texts no longer constitutes “creation” in the classical sense, the idea of authorship has indeed reached an impasse, because its relation to the work becomes fleeting and elusive: Who is the author of a book that was written written by someone else? Think Borges’s Pierre Menard.

One contributor, Scott Myles, offered a laconic: “Without the artist there’s no art.” This approach, of course, settles the matter. Just as all bachelors are unmarried, all art is made by artists. What is the difference between bachelors and art? Maybe that art and artists do not exist in a tautological relation to each other but can easily be considered as separate, because they are historically variable quantities. “Artist” was never an objective concept; the artist’s spheres of duty and influence have continually changed from the Renaissance to Romanticism, Naturalism, classical Modernism, and Postmodernism from artisan to genius to scientist to discursive hub; likewise with “art.” Appropriation and conceptual literature are only the latest turn of the screw.

What is interesting about the responses collected in the volume is that most of them ignore this context. Instead, they focus, for better or worse, on a different but related literary genre: generative code literature, which creates text through scripted algorithms. On the one hand, this shows the proximity of code and concept, that is, the fact that codes formalize concepts and make them executable; on the other hand, it indeed seemed to indicate what Tobias Roth recently claimed in Logbuch Suhrkamp, namely a “strange freneticism” gripping authors at the idea “that the machines might take the work out of their hands.” At times, this fear of the digital comes close to a new Luddism, which preferably wants to prohibit all technical interventions into the activities of the author.

Roth, too, is skeptical: “Is the artist necessary for making art today?” appears to him as “implying a situation that is not the case.” But by withholding what exactly is implied here (after all, it’s just a question), he, too, redirects the conversation without any comment toward the topic of technologically generated art. He begins, of course, with a certain nuance: It may be that the production of art can be decoupled almost completely from the author-subject – but only almost. Someone always has to make a start, provide the first impetus, which sets in motion whatever it may be. And yet, Roth immediate goes on to say: “Medium, form, tool are placed in the middle and provide neither material nor initiative; neither do they have a memory, at best a hard drive.” There are two arguments here: One that is causal and asks for the initial act, and one that focuses more on the extent in which the mean, the “tool boxes” carry out their work when it already has begun – a work that might change from quantity to quality. As it turns out, a whole human-aesthetic ontology is at play here, because Roth is concerned with the very big picture: “Which faculty, which concept signifies this gap that separates man from machine?” The answer follows promptly: “Mind (Geist) would be a name for this abyss.”

What we have here is a conjunction of the concepts “machine,” “artist,” and “mind” (or, since Geist is one of those almost untranslatable German words, spirit) – all of which come down to another concept, that of creativity. Machines (and codes are understood as machines here) cannot be ‘creative.’ As much is couched in the question of “whether machines can make music and not only play it. Or literature, or painting, it is an open-ended list.” Because machines lack mind, which is a prerequisite for creativity, and, in the end, ‘genius’, the machine will never be a producer of art; creative or ingenious is only the programmer that sets it in motion. But does it have to be “mind,” does it have to be this “mind”? Is that a good description of “mind”? Or of “machine”?

If we mean by “mind” what is meant by the now classic branch of the philosophy of mind, that is, the emergence of consciousness, the question would be: How can machines without consciousness produce art? Understood this way, machines that have no consciousness have no mind and are incapable of creation. If Geist – which, again, includes the meaning “spirit” – signifies even something like anima or spiritus, the divine breath, then all questions have to end anyway; and even the Romantics’ idea of “spirit” as genius is, at its core, theological. But what if we understood “Geist” differently? What if we conceived of “mind” in a way that blurs the difference between humans and machines?

Here, Andy Clark and David Chalmers’s idea of the extended mind might be helpful. It is a conception of “mind” that cannot be located neuronally, physiologically in the brain but which, in order to function, has to access the world, so that, in the end, genius is not a matter of “skin and skull” but takes places in the space between head, hand and, for all I care, goose quill. After all, unless you are a savant, you do not write a text in one go but produce it in a constant feedback operation between material world and consciousness. And if I am unable to write a text without seeing it on the screen or on paper, then, according to this “active externalism,” we must also locate “mind” in the paper, in the keyboard, the pen or the screen.

Nor is the word “machine” understood properly here. The metaphor of the “writing-machine” an author builds surfaces frequently in discussions about code literature. This often leads to suspicions about authorial fantasies of omnipotence and a fear that there is some will to power in the machines and the oft-invoked “algorithms.” This could be called the authoritarian fallacy. The authoritarian fallacy consists of confusing the determination of rules for the generation of text with the determination of the text itself. What this assumption omits is that such rule determination – upon which every “machine,” every “algorithm” is based on – often happens without a clear idea of its result. When it comes to literature, rules are often the opposite of authorial power and rather constitute a type of surrendering. Compared to this, classical authorship is founded much more on control.

For example, I built a program that wrote the novel Durchschnitt (Average) for me. The script, written in Python, consists of a handful of lines and operates according to an almost laughably simple recipe – it calculates the average sentence length of all the novels in critic Marcel Reich-Ranicki’s collection Der Kanon (think Harold Bloom’s Western Canon, only in German), takes these sentences and sorts them alphabetically – but the outcome is one I would never have been able to imagine, nor realistically produce by hand. My mind has jumped into the machine.

Taken together, the extended mind and the machine that knows more than me make it harder to identify tool and maker. At the very least, it becomes questionable whether we can see the machine as just another extension of human beings as “prosthetic Gods,” as Freud had it. From my primitive program, this relation can be extended ever further, so that with the increase in complexity the dependence of the program on the programmer decreases; it would be appealing to test the limit of this development. Christian Bök has noted that networked communication produces more and more language that is only exchanged between computers without ever finding a human reader. Is a text disqualified from being art if no one reads it? Dave Jhave Johnston even speaks of an “aesthetic animism” – human-machine communication is “spirited,” one could say, full of “mind.”

But maybe it would be better to do away altogether with such concepts as mind and creativity when it comes to art and literature; the mind/spirit Roth is talking about is still closest to the concept of genius, the great, autonomous, creative and powerful. It seems that romantic assumptions are at work here, which still rule the notion of author and artists, even if they could be reduced to the operational commands of calling “start” and “stop.”

Ultimately, my invitation to the writers of today is not to rest on a single, heroic-romantic concept of authorship, but rather to produce without control, to “let it write” – and by that I mean: to learn how to program. First, as coder and poet Jörg Piringer has written, to offer a practical alternative to the perpetually unsuccessful Luddism: “the poets of coming years will not sit back and leave control over the language of algorithms to companies.” And secondly, to extend the mind, which already rests within our machines, even further – and with it, the limits of literature.