Handmade images print version

What does it mean for an image to be made by hand? In general usage, calling something handmade places a certain amount of care and responsibility with the maker. If you made a pie by hand you can take credit for its deliciousness. You wave off compliments if it’s store-bought. But just as handmade pies lean on the qualities of their ingredients, even the most handmade of images are inseparable from the physical, non-human properties of the tools and materials that someone’s hands used in the making. How much we let those properties matter in our reading of an image depends on where we locate the “art” of the process; i.e. where we expect the presence of an artist doing art in a way that matters to the image.

If a graphic artist today wants to reproduce the look of an early modern propaganda poster, they might seek out digital brushes that reproduce the fine grain of a lithograph crayon or a pastel. If we are to say that an image using real crayons or pastels is more handmade than that image, we locate some of the art in the physical materials themselves. If we say the digital image they make with those brushes is meaningfully more “handmade” than one made with “AI”, then we locate the art more in the movement of the artist’s hand than in the materials.

The advent of image generating systems and other things we call “AI” has made the question of what constitutes a handmade image feel urgent enough for artist and non-artists alike to start fighting online about it. I think the mainstreaming of digital painting tools, meant to simulate specific materials and gestures, warranted the widespread public negotiation we’re now having, so I think we’ve got some catching up to do. If we are going to come up with a useful distinction between types of physical authorship, focusing on the tools and materials is a dead-end, since we can shift where we locate the art to qualify or disqualify images as we like. We can never say the hand makes an image alone, because even handprints on a cave wall benefit from the assistance of indifferent, powerfully-pigmented material. Maybe we could look to sign language, martial arts, or saman (“the dance of a thousand hands”¹) to be absolutists about art of the hand, but how useful would that be to us in the gallery or the studio?

On the other hand (ha), there are plenty of handmade images that would rather you forget where they came from. Much of the training of graphic design before the digital age was in how to present an image to a viewer that remains an image, without any reminder of the process that put it there. Then digital tools were created to remove the hand from the image even further, replacing the designer’s drafting table, Letraset sheets, and pots of gouache . But now “AI” image generators and processors threaten the jobs of photo retouchers and matte painters. If we are going to recognize the difference between the work of a photo retoucher using Adobe Photoshop before and after the addition of tools like “Generative Fill,” we must hunt for the presence of a human hand carefully and intentionally hidden by many layers of “conventional” automation.

Instead of drawing and redrawing boundaries between human and computer, we could distinguish between images in general and images that, when we look at them, remind us that they were made, revealing to some degree how they were made and who made them. We could distinguish images based on the activity of the artist’s thoughts, not their hands. Is it an image that a human thought about making, and, in doing so, did they think about how other humans will receive it?

I think this distinction is much more important and useful than how the image was constructed because, as receivers – not even viewers – of images, we also make (and remake) images in our minds. This distinction also allows us to move towards a more inclusive term that doesn’t assume everyone has the same kind of body.

The entanglement of thought and seen images

I listen to music, stare out the window, and nap in the studio because there is something concretely useful to me as a painter that comes out of the imagining – or really, imaging that comes from music or dreams (day- or otherwise). It’s not inspiration to paint, it’s the same work² as painting in a fundamental way. Not in every way, of course. I paint with oil paint because the physical qualities of the paint turn the process of making a picture with paint into a collaboration with material in a way that is significant to me and the pictures I make. The gap between imagination and the canvas may be massive and charged with electricity, but your mind is one of the contact points between which the creative spark sparks. Even if you have no idea what you’re going to make before you pick up a brush and try to cross that gap, you produce mental images of some kind – many kinds, really – in tandem with the physical image. Whether you need paint, dirt, or pixels to create an image, we cannot deny the importance or the existence of the thought image, the medium-less image.

The types of handmade images that I am concerned with as a painter are the ones which requires their makers to remain aware, present, and curious about both kinds of images that constitute the life of a picture. Ones that can only be made with an understanding that pictures exist in an entanglement between thought images and images fixed in a medium. I think this sort of image should be of equal concern to anyone interested in image generators or “AI” in general.

The artist Vija Celmins declared a similar interest, in

…that kind of double reality, where there’s an image, but the image is here in another form, and when you look at the work, you have that double thing you should have all the time, where you’re looking at the making. A kind of redescribing of the surface, and the image is interwoven with that surface.³

I wanted… an image that was in your mind, that was vast, and the reality which was very restrained and flat and made, and that was actual.”

Celmins makes a particular type of hyper-realistic art which I wouldn’t call photorealism, because it seems more concerned with the object than with the photograph. She makes sculptures of stones, drawings of the ocean, paintings of space. By referring to them as “redescriptions,” she implies that the object itself, before she found it or looked at it or drew it, was first described by its own existence. So it follows that art that re-describes it should be fully entangled with its existence, not simply its image, not just the light reflected off its surface.

If you go to art school, you enter a context where this is expected of pictures. You leave with a shorthand for it – interesting pictures are “rigorous”, or maybe “serious”. They are interesting because when they come alive in your mind, they don’t stop moving. They can’t be pinned down, they feel like visitors you can talk to rather than artifacts that can be owned or consumed.

If you set about trying to make and understand images like this you can work on it for most of your life like Celmins has. But you will not be stuck in your mind, or in the picture, you will be firmly pressed against parts of the outside world, examining them and your experience of them to an exacting level of detail and depth.

Celmins says:

There are millions of decisions of course when you’re working on something, so I try to be alive and present for that moment. It’s a kind of record of moments of attention.

Pictures that are “made by hand” in this way are the results of a process of mental imaging and picture making. The hands are often just a way of touching the world you are thinking of. We might be better off calling this kind of image “mindmade”.

Ignoring the spirochetes

If this is the kind of art you want to have and to make, then it should be obvious that machines aren’t close to making it any time soon. Computers got so good at making pictures so fast that we forgot – or were encouraged to forget – how bad they are at thinking about pictures. Machines do not look. We have been too ready to contribute our minds’ images to breathe life into images with no mind behind them.

It’s natural that we would look at a complex picture, and start asking questions about the amount and types of thought that went into making it. That’s what being an art student or a “rigorous” artist is all about. I’d venture that it’s somewhat less natural that, in doing so, we would center our conversations around technology, technologists, and the businesses and products they make. Talking about art, artfulness, and consciousness in these terms is ultimately pointless. It is a political act, akin to firing your AI ethicists⁴, to choose not to seriously engage the voices of all the many artists, writers, scientists, and non-technologist-thinkers that have been asking these same questions for much, much longer than the PR department at Google has. This is why I think we should meet these tools as art students, because otherwise we’re just customers.

I’m also hoping that, if we do, we will understand more deeply what’s at stake in the attempt to make machines that could make art. Or rather, machines that could be art students with us. Right now, ChatGPT can pass the Bar, but it can’t get an MFA. To graduate art school, it would have to show up, ask questions, keep showing up, keep asking questions, and make something, not even an image, that was intrinsically tied to that process. It would have to be in the world with us.

Perhaps I keep laboring this point because I’ve only read one compelling description of an artificial intelligence that acted like that, and I think we should put more effort into imagining such a thing. It was in Lewis Thomas’s essay An Earnest Proposal⁵and the AI he described ends up saving humans from destroying themselves, even though we asked it to launch nuclear missiles.

Lewis Thomas published An Earnest Proposal in response to the growing anxiety over automated computer systems all the way back in 1971. He addresses the logical concern that the “practical men” running our countries would enlist intelligent computers, if we had them, in the complex decisions of nuclear warfare, in deciding acceptable levels of “megadeath”. Like today’s AI alarmists, good or bad faith though they may be, Thomas wanted the deciders to slow down a little.

I have an earnest proposal to make. I suggest that we defer further action until we have acquired a really complete set of information concerning at least one living thing. Then, at least, we shall be able to claim that we know what we are doing. The delay might take a decade; let us say a decade. We and the other nations might set it as an objective of international, collaborative science to achieve a complete understanding of a single form of life. When this is done, and the information programmed into all our computers, I for one would be willing to take my chances.

He nominates Myxotricha paradoxa, a protazoa that lives in the digestive tract of Australian termites.

It is not as though we would be starting from scratch. We have a fair amount of information about this creature already – not enough to understand him, of course, but enough to inform us that he means something, perhaps a great deal.

It turns out termites don’t digest wood, Myxotricha paradoxa does it for them. It poops out lignin, which the termites then use to build their metropolises. Myxotricha uses flagellae to swim around inside termites in search of wood. But, oops, they’re not flagellae, they’re spirochetes – which are also basically separate, living things. And – oops again – some organelles, which are actually their own bacteria again, live at the connection point between spirochete and Myxotricha. My knowledge of what they’re doing there is stuck at 1971 pop-science levels which is “probably contributing enzymes that break down the cellulose.” The description, unsurprisingly, goes on and gets more complicated.

So, Thomas says, fine, build computers that tell us how and when to blow up the world. Just make sure they understand one single tiny part of it first. Thomas ends with this:

I take it on faith that computers, although lacking souls, are possessed of a kind of intelligence. At the end of the decade, therefore, I am willing to predict that the feeding in of all the information then available will result, after a few seconds of whirring, in something like the following message, neatly and speedily printed out:

Request more data. How are spirochetes attached? Do not fire.