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

General definition

The central question a doctoral project sets out to answer, from which its subsidiary questions, methods and outcomes follow. It defines the scope of the research and the kind of knowledge the project claims to contribute.

Relation to the practice

‘What form of creative intelligence surfaces when artists and AI agents create within the rule-based instruments of the MIDI era, preserving digital heritage through AI-extended active use, and what makes this practice transferable to other practitioners?’

The research investigates the question through its own practice and through work with other practitioners. Five subsidiary questions divide it: design rationale, shared models and overload, agency and authorship, transferability and visual music. The research stated the question in September 2026, and it is still under revision.

Question in the plan

The question in the plan is how artificial intelligence can mediate between legacy creative systems and contemporary practice. It states that this mediation preserves creative intelligence and enables new forms of audiovisual expression. The wording quoted above is the later formulation. The later wording moves the emphasis away from what the model does between the systems, and the question now is what forms of creative intelligence appear when an artist and agents work inside them.

Both questions ask for the same work, and CONDUCTOR and the works now do that mediating work. The plan addressed the second half of the current question, what makes the practice transferable, through educational partnerships and adoption. That assessment reads the trajectory against the earlier question.

Early question

The research worked hardest on the question in the first months of 2026. It set two formulations against each other. The research chose the question of what new forms of expression and interaction emerge when artificial intelligence is embedded within the constraints of endangered legacy creative systems. The word endangered replaced obsolete because it names both the cause and the stakes.

A later session in the same month reduced the question to what becomes possible when contemporary artificial intelligence is constrained by historical systems for creative interaction. The session also put the question of what that constraint reveals about creative technologies in 2026. The research drafted four subsidiary questions at that stage, one for each of the doctorate’s four researcher roles, and it removed terms that pre-loaded the answer, collaborative intelligence among them.

In April the research reworked the question again, and it moved towards how the roles of human and artificial intelligence are distributed in creative practice. The April wording also put the question of what the early intelligent music systems designed for human-machine interaction can contribute to the answer. This time the research also removed the either-or framing of the two roles. In May the research put the question in Dutch for the supervisory committee. The Dutch wording put the question of how contemporary artificial intelligence can continue the human-centred design philosophies of the historical creative programming languages, and what that means for human-machine collaboration in creative practice.

At the end of May the research examined the question in English and named three problems. The first problem is that the question never states the baseline it compares against. Second, it does not distinguish legacy systems from constrained ones. Third, nothing yet joins the work of porting a language to the compositional material that would answer the question.

In July the research tested the question against what makes a research question finished. The test requires that, in practice-based research, the practice makes part of the argument and does not only illustrate it. After two objections from Gardner, the research revised the question again. The first objection was that early intelligent music systems leaves out lighting and other distributed intelligence. The second objection was that the phrase ‘contribute to the answer’ reads as too casual, and the question in the wording that survived is what earlier systems for human-machine creative interaction can reveal.

In September 2026 the research reached the current wording. An account of that month states that the question crystallised over eighty exchanges. Gardner kept one reservation, that the term ‘rule-based’ sits uneasily with the criterion the corpus has to meet. A comparison of the plan of 2025 with the practice of 2026, made in October, shows the question moving from autonomous human-machine co-creation to a more limited arrangement that keeps the human as the anchor. The change affects the research’s position as well as the wording. The current wording still contains the tension Gardner named, and it remains under revision.

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.