Writing sucks (except when you do it)

Here I am, typing away at my computer. I’m working on a paper, building an argument I’ve thought about for a long time. It’s not easy, but after a while things begin to flow and sometimes I surprise myself with what I wrote. I read my last couple of paragraphs, and suddenly begin to see a new point that’s beginning to emerge. I had no idea that this was in there, but now that I’ve seen it, it makes perfect sense. Of course, this is what I wanted to say! I feel like a creative genius.

Here I am, cooking what I hope will become a nice Christmas dinner. I’ve used this recipe once or twice before, but I keep checking the book just to be sure. When I open my fridge to get a jar of preserved lemons, I see a leftover pack of coconut milk and realize that this is an extra ingredient I can use to top things off. It’s not in the recipe, but now that I’ve seen it, I know that it will work. In fact, why wasn’t it in the recipe in the first place?!

Here I am, preparing a lecture for my students on different models in organizational psychology. This is material I’ve been covering for years, but every year I like to update my slides a bit. Looking over what I have, I realize that there is a way to frame these models that I’ve never used or even seen before – one that should make the core points easier for students to follow, and more interesting for me to present. With renewed energy, I continue working on my slides. This lecture will be more fun than I thought.

 

All these examples reflect a creative process. We often envision this as a progression from raw idea to final product. Things start with a vague notion somewhere in the mind, and then move through various forms of elaboration to something like actual output, which could be a physical product, an improved procedure, a written paper, a meal, a painting – anything, really. Psychologists typically define creativity as ‘the ability to produce work that is both novel and useful’, and that can happen in almost any domain or activity. Of course, this is part of why it’s such a fun topic to study.

 

Apparently, creativity revolves around producing work, delivering some kind of output.

 

What is usually emphasized about this definition, is the combination of novelty and usefulness –creative ideas or products are different from what’s already there, but also somehow work or fit within a particular context. This makes sense, but there’s something else in this definition, that has tended to escape notice ­– perhaps because it seems so obvious. Apparently, creativity revolves around producing work, delivering some kind of output. Again, this makes sense: what is creativity for, if not for making creative things? Especially if we want to reap the benefits of creativity and innovation, we have to be able to observe and measure it. This is pretty straightforward behavioral science, and the basis of the kind of research I’ve been doing for decades. Then again, in Psychology things usually aren’t as straightforward as they seem, and creativity is no exception.

Product or process

About a year ago, when I taught my master course on creativity and innovation, one of my students sent me an interesting chapter she’d come across. The chapter, written by Julia Langkau, argues that we can look at creativity by focusing either on the output (a product) or on the mental states people go through when doing creative work (a process). Her point is that the former has received by far the most attention, but the latter is no less valid an approach. I found it a very interesting read, but it didn’t immediately strike a creative spark; I had no urgent use for it at that moment. But then something changed. ChatGPT and other forms of generative AI became more and more prominent (and better at producing novel and useful output), and suddenly everyone was asking: is this the end for us? Can an AI finally be really creative? Are humans going to be obsolete?

At first I found those questions merely boring; after a while I began to find them very annoying. Something was wrong here, I felt; people were asking the wrong questions and worrying about the wrong problems. Of course, we should worry about job security, especially for those in precarious and undervalued positions. But I felt the issue was a deeper one – and then I realized that this was because all of these discussions revolved around creativity as output. The question whether ‘an AI can be creative’ was reduced to the question whether ‘an AI can produce work that we find creative’. Where was the psychology? Whatever happened to the process?

Early pioneers of artificial intelligence and creative cognition, like Douglas Hofstadter and Margaret Boden, were interested in creating computational models of the way we think and come up with new ideas, such as our ability to make and recognize analogies. They were not looking to create huge and energy-intensive models that could emulate human output by dint of sheer computational power; I’m pretty sure they would have found that cheating, and –worse­– uninteresting. They were trying to understand, and model, what we do when we think creatively.

 

And now even my favorite kind of behavior, creativity, was becoming dehumanized.

 

As the publicity around AI got whipped up into a positive frenzy, I increasingly found myself whispering things like ‘the ability to produce work that is novel and useful is insignificant next to the power of true creativity’. Obviously I couldn’t argue with the performative power of AI, but I wasn’t at all convinced that this reflected ‘creativity’ in an interesting or important sense. In fact, the whole discussion seemed rather problematic to me. The problem, as I came to see it, was not that we were anthropomorphizing AI, turning machines into humans. Everybody loves a cute robot, after all. The problem was that we had fallen back into an old habit of seeing people as a mere means for producing output, turning humans into machines. And now even my favorite kind of behavior, creativity, was becoming dehumanized.

Outsourcing creativity

It hasn’t taken long for AI to find its way into the way we study and teach. Students, it is often argued, will work with AI regardless of what we think – either now, during their studies, or later, as professionals. If writing a high-quality text is a redundant skill that can be outsourced to a machine, it no longer makes sense to want to teach or assess it. What we should focus on, instead, is teaching students the crucial academic skills that are at the heart of our profession: identifying interesting problems, processing literature, asking smart questions, making a sound argument and generating interesting new ideas. Writing, on the other hand, is just a matter of producing output, and a machine can do that.

 

The dichotomy between written output and learning process is a false one

 

The positive take on this is that we might get a chance to finally go back to our core business: teaching people to develop interesting thoughts and theories, and to adequately test them in a way that will produce useful and valuable answers. Not having to worry about students’ involvement (or lack of involvement) in writing a text could be liberating! But the more I think about it, and the more I reflect on my own experiences, the more I begin to realize that this separation between output and process is false. The two always go together – that is, in humans. Of course, writing will usually produce some kind of output. But it would be a mistake to consider writing an exclusively output-oriented activity. The point of writing is not the final text; at least, not for the author. The point is to go through the process of writing, of finding out what you want to say, and of learning something in the process. I think this is true even for experienced authors, but definitely for our students. We can think deeply about our theory or research, and things may look crystal clear to our mind’s eye, but it’s only when dealing with actual words, sentences and paragraphs that we’ll find things weren’t as clear as they seemed to be.

This, I think, has an important implication for the role of writing as a learning tool: the dichotomy between written output and learning process is a false one. An essay is not called an essay for nothing – writing often is a matter of trying things out to see how well they work. In other words, the act of writing is what helps us go through the learning process. Outsource that, and there’s not much left.

The not-quite-fictional three examples I’ve given at the beginning of this text illustrate what I have in mind here. The creativity of writing, cooking, teaching, is in the making. If we deny ourselves the process of making something, the annoying complications, the beautiful mistakes, the unexpected insights – if we deny ourselves all of that, we deny our own potential for creativity and development. Thinking is not some disembodied act of pure cognition, and writing is not a matter of simply putting down what you already have in mind. I do some of my best thinking when I’m trying to articulate the vague thoughts I have been carrying around in my head, and I often learn most from not (yet) being able to do so. It’s difficult and frustrating, to be sure – but nowhere near as frustrating as seeing a machine do something I would have benefited from doing myself.

 

Image credit: photograph by Ami Harikoshi; licensed with CC BY 2.0

 

References

Langkau, J. (2022). Two notions of creativity. In C. C. Pfisterer, N. Ratgeb, & E. Schmidt (Eds.), Wittgenstein and Beyond (pp. 254-271). London: Routledge.