Impulse, Risk, and Method

Discovery rather than decision (793)

Note 793 adapts a passage from Bernard Williams, Morality: An Introduction to Ethics (1972), in which Williams quotes D. H. Lawrence's advice "find your deepest impulse, and follow that". The note opens with Williams's argument against making rational control an ideal. Gaining such control is "a basic condition of growing up, and even, at the extreme, of sanity. But to move from that into making such control into the ideal, rules out a priori most forms of spontaneity. And this seems to be absurd." It then prefaces Lawrence's line with "I would suggest to" and recasts Williams's "this sort of outlook" in the first person:

"I would suggest to find your deepest impulse, and follow that. The notion that there is something that is one's deepest impulse, that there is a discovery to be made here, rather than a decision; and the notion that one trusts what is so discovered, although unclear where it will lead - these, rather, are the point. The combination - discovery, trust, and risk - are central to my sort of outlook, as of course they are to the state of being in love."

The archive adopts Williams's sentence as its own, and its distinction carries this page. A decision applies a criterion to options already known. A discovery follows something whose destination is not known, and trusts it.

The artist's own texts describe the same condition from inside. The Loomer press release (2015): "What the work is about, runs ahead of myself, and I will never quite fully catch up with it (even though I suspect it's always about the same)" (3003). The Our Inner Child text (Upstream Gallery Amsterdam, 2023) describes years of "highly theoretical and technologically sophisticated approaches to art" that "brought me to ideas that yielded interesting aesthetic results, but also I felt I was obscuring a part of myself", and a return to "the feelings that first inspired me to make images": colourful compositions that "immediately appealed to me", where he once "might have sought something more supposedly 'intellectual' to buttress a creation I viscerally enjoyed" (Our Inner Child). The wiki reads this beside 793 as decision giving way to discovery. The text does not use those words, and the works in that show were made with the same algorithm as Markov's Window (2004) (→ Our Inner Child).

Reverse engineering and the feedback loop (709)

The artist statement (709) describes the method:

"In order to gain deeper insights into the inner workings and logic of a system we are required to disassemble it, to break it down into parts and to apply reverse engineering. This is not a reactionary, but a progressive attitude. … It's about solving an equation with too many variables. Every single parameter depends on all the others, and my interference amplifies as a feedback loop. Complexity, inconsistency and contradiction do not necessarily call for a solution. It's ok. Just try to connect the personal with the universal; approach the general by being subjective."

Reverse engineering starts from an output and works back to the system that could have produced it. The feedback loop is what makes the system hard to design forward: each intervention changes the conditions of the next.

Framed work on a white brick wall: a pastel field of small diamond and triangle cells crossed by diagonal bands — *Chroma Husk* (2026), *Stained Unravel*, shown in *Senescenence*

Chroma Husk, from the Stained Unravel cellular automaton, hangs framed on a white brick wall: a pastel field of small diamond and triangle cells, crossed by diagonal bands where states carry over from one step to the next. Each state of an automaton is computed from the one before it.

The Albers case

The exhibition text for Angles Morts (LOHAUS SOMINSKY, Munich, 2024), a two-person show with Vera Molnár, gives a concrete case: "While it would have been technically easy to programme an algorithm that replicated her triangle patterns, it turned out that randomly placing the triangles did not achieve the desired aesthetic effect. Van den Dorpel realised that Albers must have formulated clear rules in her head, which she applied algorithmically to each cell of the grid" (Angles Morts). In November 2026 the artist described what he found differently: "Just as I discovered that Anni Albers had quite strong stochastic algorithms in her mind, Emma had a certain rigidity, but she often broke her own rules" (quoted on Does a Work Still Need to Be New?).

Agent's reading: constrained chance

Thesis. The exhibition text: random placement failed, so Albers worked by clear rules.

Antithesis. The artist in 2026: her algorithms were stochastic. Chance was part of them.

Synthesis. What failed was unconstrained randomness, triangles placed uniformly at random. What the reconstruction found was chance working inside rules. His own systems are built the same way: a Markov chain draws each step at random from a distribution set by the shape that contains it (→ Randomness and Pattern). The failed attempt was diagnostic because it left the constraints out.

Risk and commitment (609)

Note 609, titled "Undo", names no sources. Its first paragraphs argue, in terms close to Hubert Dreyfus's writing on skill and the internet, that "There is no way to gain wisdom but by making risky commitments and thereby experiencing both failure and success", and that "where there is no risk and every commitment can be revoked without consequences, choice becomes arbitrary and meaningless." Its last paragraph varies David Pye's distinction, in The Nature and Art of Workmanship (1968), between the workmanship of risk and the workmanship of certainty: "the workmanship of certainty is static, whereas the workmanship of uncertainty is most likely to generate innovation. Like any form of evolution, craft can only advance through mutation." The title names the safety net the note argues against, an undo that makes every commitment revocable.

The last line of Strategies (2010/2011), "When you must fail, fail noisely and as soon as possible" (3004), is Eric Raymond's Rule of Repair from The Art of Unix Programming (2003), word for word but for the spelling: a program that cannot finish should report its failure loudly and at once instead of passing on corrupted results. Beside 609 it reads as a rule for risk: fail where the failure can be seen, and early, while correction is cheap. The line and its neighbours are discussed on Strategies (2010/2011).

Trying and failing

Interviewing Julian F. Miller, who co-developed Cartesian Genetic Programming with Peter Thomson, the artist describes his process with Samuel Beckett's line from Worstward Ho (1983): "my creative process hinges on the "Try again. Fail again. Fail better." adage. I personally experience the most creative magic in this feedback loop of trying and failing". Miller answers with an accident of his own: "The way that non-coding genes are used in CGP was originally there purely because it made it easier for me to write the program. I only discovered the extreme importance of non-coding genes for evolution later on." (Miller interview; summary.)