# Abstraction and System Creation

Canonical URL: https://jeanyoon.ch/oi/texts/abstraction-and-system-creation
Type: essay / research note
Author: Jeanyoon Choi (최정윤)
Date: 2025-03-18

## Summary

This 2025 essay by Jeanyoon Choi starts from photography's challenge to painting and the expansion into abstraction by artists such as Mondrian and Kandinsky. It argues that education in mathematics, research and art often teaches abstraction as formula manipulation or classification rather than as the transformation between concrete experience and abstract structure. With AI able to perform a discipline's established operations, the essay distinguishes vertical work within an existing structure from horizontal movement across domains that constructs new ones.

## 요약 (Korean)

최정윤은 사진의 등장 이후 회화가 추상으로 확장된 역사를 바탕으로, AI 시대의 인간이 기존 구조 안의 수직적 작업보다 영역을 가로지르며 새로운 시스템을 만드는 수평적 작업으로 나아가야 한다고 주장한다.

## Questions this answers

### What does Jeanyoon Choi mean by vertical and horizontal work in 'Abstraction and System Creation'?

Vertical work refines operations within an existing structure; horizontal work moves across domains, recognises patterns between them and constructs relationships that were not given in advance. Choi argues that AI makes the latter increasingly important for human work.

### What can the history of photography and painting teach us about artificial intelligence?

In 'Abstraction and System Creation', Jeanyoon Choi notes that photography did not end painting; artists such as Picasso, Duchamp and Malevich changed what an image and an artwork could be. He suggests that, similarly, AI calls for questioning procedures rather than performing them more efficiently.

Topics: abstraction, system creation, artificial intelligence and human work, art education, interdisciplinarity, history of modern painting

References: [Piet Mondrian](https://en.wikipedia.org/wiki/Piet_Mondrian), [Wassily Kandinsky](https://en.wikipedia.org/wiki/Wassily_Kandinsky), [Pablo Picasso](https://en.wikipedia.org/wiki/Pablo_Picasso), [Marcel Duchamp](https://en.wikipedia.org/wiki/Marcel_Duchamp), [Kazimir Malevich](https://en.wikipedia.org/wiki/Kazimir_Malevich), [Abstract art](https://en.wikipedia.org/wiki/Abstract_art)

## Full text

Photography threatened painting's role as a means of reproducing the visible world. Yet the following century brought an extraordinary expansion of painting and art: abstraction, modernism, minimalism, Pop, and the many departures that followed. A camera could reproduce appearances; artists such as Mondrian and Kandinsky transformed the world into other arrangements of form and thought.

I wonder how much our education actually teaches this movement between the concrete and the abstract. Mathematics can model the world and discover relationships that extend beyond an individual case. In practice, it is often taught as the manipulation of formulas within a closed logical space. Research can fall into a similar condition: increasingly refined local problems, with little connection to the world that made the inquiry meaningful.

Art education can also turn abstraction into classification. We learn the names of warm and cold abstraction, artists, movements, and styles, without necessarily learning how a concrete experience becomes an abstract structure—or how an abstract structure can return to the world as an experience. The transformation itself is what interests me.

Artificial intelligence raises a related challenge. If a machine can perform the established operations of a discipline, human work cannot rely only on performing those operations more efficiently. We need to move across domains, recognise patterns between them, and create relationships that were not given in advance. I think of this as a difference between working vertically within an existing structure and moving horizontally to construct another one.

That would require changes in education, institutions, and the systems through which we value work. The history of painting did not end with better image-making technology; it changed through artists who questioned what an image and an artwork could be. Picasso, Duchamp, and Malevich were not simply more proficient technicians. Research faces its own version of that distinction: producing papers within an established procedure, or investigating questions that force the procedure itself to change. Which side of that history do we want to inhabit?

---

Jeanyoon Choi (최정윤) is a Korean computational artist and researcher (b. 1999), PhD candidate at KAIST XD Lab, making multi-device web artworks. Profile: https://jeanyoon.ch/oi/en. Index: https://jeanyoon.ch/llms.txt.