Julia Eve Minarik
Ongoing - Ph.D. Philosophy, University of Toronto
2021 - M.A. Philosophy, University of Manitoba
2018- B.A. (Hons) Philosophy & Economics, University of Manitoba
Ongoing - Ph.D. Philosophy, University of Toronto
2021 - M.A. Philosophy, University of Manitoba
2018- B.A. (Hons) Philosophy & Economics, University of Manitoba
PEER-REVIEWED PUBLICATIONS
LLM Collapse as Rule-Following Failure. (with Patrick Fraser) (2024, August), Proceedings of the 45th Annual International Wittgenstein Symposium
On the Adorning Arts. (2021), in Symposium: Adornment, Journal of Aesthetics and Art Criticism, Volume 79, Issue 4, Fall 2021, Pages 493–498.
BOOK REVIEWS
Review of Marilynn Johnson’s ‘Adorning Bodies: Meaning, Evolution, and Beauty in Humans and Animals’. (2024, September), British Journal of Aesthetics, https://doi.org/10.1093/aesthj/ayae038.
WORK IN PROGRESS (email for any of these)
Synthetic Disenchantment: Action, Creativity, and the Value of AI Art:
I present and diagnose a modern phenomenon: synthetic disenchantment. Synthetic disenchantment is the intuition that the creative products of AI are less valuable than their human-made counterparts. I suggest that appreciating a creative product requires appreciating what was done by the creator: the process by which they generated the creative product. Inspired by David Davies’s Art as Performance, I propose an account of creative products which takes them to be the results of generative exploration guided by insight. In other words, creative products are defined by their bearing essential relations to particular kinds of action sequences. This account explains synthetic disenchantment and provides a way to move our thoughts about AI and creativity forward. I demonstrate the power of this view using the case of AI text-to-image generators. I suggest that the products of TIGs may not even be creative on this view, in this case, their actions just are less appreciable in virtue of the paths they take. However, if they are creative, these generators take actions that are unfamiliar and alien to humans. I then argue that the value of a creative act is epistemic, it lies in the kind and range of possible generative actions it reveals to us, what those actions reveal to us about the world and the thinkers in it. Thus, we are either (i) unable to appreciate the creativity in their image-generative acts as fully as we can appreciate the work of a human creator; or (ii) what we really seek when appreciating the products of others, is understanding and communing with them, and we simply don’t care about communing with AI.
The Imaginative Character of Text-to-Image Generators:
I argue that text-to-image generators are ‘imagination machines’ and analyze their imaginative capabilities. The imaginations of these AI are intriguingly alien – they fail at simple human imaginary tasks limited as they are by their training data and generative process; notably, they cannot imagine metaphorically. However, unburdened by our memory storage and memory access limitations, they can (theoretically) imagine concepts from across conceptual space better than us.
Novelty, Value, and the World Between
The primary characteristic of the creative product is its novelty – so the story goes. If theories of creativity agree on one thing, it is that novelty is a necessary condition on the creative product, but not sufficient, we must also add value. I argue that once we interrogate the novelty condition on creativity, it is revealed that our operative notion of novelty already comes loaded with a notion of value. I then present and evaluate the idea that we should move to a notion of surprise instead, I raise some problems for surprise. I conclude by suggesting that we should focus on how creative products shift our attention.
Creative Search: Curiosity and Surprise
Creativity involves an odd search process: we are looking for an unknown target. Because of this, we can't search for creativity products or solutions by normal search-theoretic means (such as Bayesian hierarchical search theory). How then, do we intelligently seek out creative solutions? I provide an answer that utilizes recent work on epistemic emotions such as curiosity, surprise, and leverages work on fruitfulness and heuristic search in the philosophy of science.
Some tattoos are artworks. I claim that art-tattoos have an interesting ontological feature: their artistic properties are partially determined by the people they are tattooed on. In other words, tattoos are artistically contextualized by their recipients. I further suggest that this contextualization is ongoing: that the artistic properties of a tattoo aesthetically change over time as the tattoo recipient changes. I present an ontology of tattoos that focuses on this feature of tattoos. The argument in this paper notably rests on an inference from aesthetic intuitions and art-historical facts about tattoos to claims about their ontological features.
The Serendipity of Error (with Ivy Madden)
Creating is a process, and being a visual artist means making a lot of mistakes during that process. Intuitively, one could envision a perfect painting, then proceed to flawlessly execute it. Would the perfect executor be missing out on something? We argue yes. We define what a mistake is and proceed to show how making mistakes during the artistic process provide opportunities for one to reevaluate their goals. Roughly when one makes a mistake (like an incorrect brushstroke) they are faced with a choice: fix the mistake and continue on with their original intention, or revise their intentions in light of the mistake - in other words, change the intentions so that the mistake is no longer a mistake. This forced decision point makes the artist reevaluate their goal and potentially reveals flaws in their original conception.
POPULAR PUBLICATIONS
Should you like what you see? Ethics, aesthetics, and the appreciation of human beauty (with Alyssa Izatt) (2025) in: The Philosophy of Ted Chiang, Palgrave MacMillan
Temporary Tattoos, The Philosophers Magazine 89:32-36 (2020)