Part of Temporal Constructs Appendix, which contains details behind the research: dates, quotations, descriptions of instruments, and sources.

General definition

A distinction between using a system to extend one’s own thinking and handing thinking over to it.

Relation to the practice

Cognitive offloading, held here through Magnusson (2009), is the other side of that difference. The pair of terms used here follows that distinction.

In depth (optional read)

Ingold (2013) writes of thinking with materials, and Clark and Chalmers (1998) describe thought extended into the things that carry it. Thinking with is the symbiotic kind. Ideas integrate until their origin recedes, and that origin remains recoverable. The artist’s capacities grow over time. Thinking for is delegation based. The origin becomes irrecoverable and the idea is not examined, or the artist’s capacities decline.

The distinction runs through the Temporal Constructs professional doctorate practice. Losing the origin of an idea can also be a sign of integration, as long as the origin remains recoverable. For facts, citations and quotations, the origin decides what must be verified.

Intermediate cases

The distinction rests on whether the origin of an idea stays recoverable, and whether the artist’s capacities grow or decline. The distinction therefore applies to an episode. The same agent in the same session can be on either side of it at different moments.

Read as two axes, the settled cases are the corners. An origin that stays recoverable with capacities growing is thinking with. An origin that becomes irrecoverable with capacities declining is thinking for. The other two corners describe the grey area.

One grey area, for example, is an origin that stays recoverable while capacities decline. The artist can still say where an idea came from, and has stopped exercising the judgement that would change it. The session reads as thinking with, while the practice behaves as thinking for.

The second grey area is an origin that becomes hard to recover while capacities grow. If the source of an idea can no longer be assigned to either human or AI, the account describes that as a sign of integration. The decision record records it as a category of its own. That category is a decision arrived at through the work itself, whose origin lies with neither the artist nor the agent.

These corners are not fixed points. The direction of travel matters more than the position. Thinking for becomes thinking with by recovery. The artist takes a decision back and reworks it so the agent’s output becomes material.

The artist learns to read what the agent did until the decision is legible. The artist rebuilds the part of the instrument the agent was standing in for. The action then becomes expressible in the artist’s own vocabulary.

Thinking with becomes thinking for by attrition. Attrition runs through the check that is not made, or through a volume of agent output larger than the artist can read. Acceptance then becomes the default, and no decision was made for it.

One route runs in both directions. An origin that recedes keeps receding. Whether that reads as integration or as loss depends on whether the artist’s capacity grew with it. This model also follows the concept that progress tends to involve loss, and the cognitive limitations of the mind.

What decides how a grey area resolves is not who produced the output. Whether the artist can say what was decided and on what grounds matters. It also matters whether a route to disagree remains that costs the work nothing. It also matters whether the artist’s attention is present at the decision or arrives after it. When the first two conditions are in place, a relation stays thinking with even when the agent produced everything.

The agent writes, and the artist’s ear remains the arbiter. When those conditions lapse, the relation has moved towards thinking for, even though the artist remains the author. These concepts also apply to authorship and disclosure.

The distinction also has a limit. When the installation worked at MU autonomously, nothing was thinking with anything, and nothing was thinking for anyone. The system was running a disposition the artist had set earlier, and it held that disposition on its own. Agent work of that kind is neither of the two modes.

The account describes that work as the artist’s earlier thinking being executed, and not as either mode continuing in the artist’s absence. The same limit applies to the agent-written code that KML transfers into Trip-A-Tron. There the artist’s part lies in judging whether the result can be relied on at all.

Stating the grey area matters, because it makes the position something the artist maintains. If the two modes were properties of a system, the thesis would be a claim about design, settled by inspection of the tools. As a relation that can be lost and recovered within a single session, the position becomes a practice. The instruments are built for that practice: Choice of involvement, the Capability gradient, legible errors and the Conversation layer. Stating the grey area also puts the evidence in the sessions. No artefact shows which direction a relation was travelling in at the moment that mattered.

Unforeseen

Here the unforeseen situations, sometimes difficult, sometimes unexpectedly interesting, will be described as the project runs. Many of these unforeseen situations have already occurred and have been recorded. They will be added in the time to come.

Evidential basis

The instruments for checking the distinction are held. An idea’s origin is meant to stay recoverable in the Decision log. Authorship is where that matters for outcomes. The recorded variations of ALTERITY are where the difference between the two modes would show. The Decision log also notes the limitation, which is that the material exists as session history and has not been turned into an account of who decided what. Until it is, the distinction is a framework the research works by, and it awaits the finding that would demonstrate it.

Framed as ethics

The distinction was carried into the methodology and ethics text written in October 2026. There the boundary it draws is stated as an ethical one and as a working one. The research makes one difference visible: thinking with an artificial intelligence and being thought for by it. The first leaves the authorship of the work where the research places it.

The following pages are part of this page but are not needed to understand it. They can be opened from here for further exploration.