Formal Sequence

In the AI Story

Kubler developed the concept during decades of work on pre-Columbian Mesoamerican art, a field where individual makers were typically anonymous and biographical attribution was impossible. The absence of biography forced a structural question: if the maker cannot be identified, what organizes the history of these objects? Kubler's answer was the problem itself. Objects were linked not by authorship but by the problems they addressed, and the problems persisted across generations of anonymous makers. The pointed corbel arch was not the expression of a single genius but a position in a sequence of solutions to the problem of spanning interior space with stone, a sequence that predated any surviving example and continued long after the last identifiable architect had died.

The formal sequence differs from related concepts in ways that matter for the AI analysis. It is not a style, which is a descriptive category applied by historians; it is a structural reality embedded in the objects themselves. It is not a tradition, which implies conscious transmission between makers; sequences can be traced by their structural logic even when the makers are unaware of earlier entries. It is not a period, which is temporally bounded; sequences can extend across periods, and multiple sequences operate within any given period. What defines a sequence is the persistence of the problem and the structural continuity of the solutions.

The sequence has a shape. It has an early phase characterized by rapid formal innovation, a middle phase of proliferation, and a late phase of diminishing formal returns approaching sequence exhaustion. Each phase has different properties for the makers who enter it. Early entrants face a landscape where formal possibilities are wide open and each solution generates new problems. Late entrants face a landscape largely mapped, where remaining moves are refinements of possibilities already demonstrated. The shape of the sequence determines more about what a maker can accomplish than personal talent or training — the structural fact Kubler called entrance.

AI interacts with formal sequences as a participant of a new kind. Trained on the accumulated outputs of existing sequences, it can generate artifacts that occupy positions within those sequences with extraordinary fluency. What it has not demonstrated is the capacity to perceive that existing sequences are insufficient — to recognize the structural absence that prompts the opening of a new sequence. The formal sequence, as Kubler's unit of analysis, reveals what AI can and cannot do with more precision than any alternative framework, because it separates the question of position within a sequence from the question of who produced the position.

Origin

Kubler introduced the formal sequence in The Shape of Time (1962) as a replacement for the biological metaphors that had organized art history. The life-cycle vocabulary — birth, maturity, decline, death applied to styles and periods — had become so pervasive that no one noticed it was a metaphor. Kubler replaced it with a vocabulary drawn from signal theory and electrodynamics: chains of linked solutions, impulses, relay points, increments and losses in transit. The formal sequence was the foundational unit in this new vocabulary, and its structural character — independent of biography, independent of biology — was what made it durable.

Key Ideas

Problem-persistence defines the sequence. What holds a sequence together is not style or period but the persistence of the problem its solutions address.

Structural position, not authorship, is the primary property. Where an artifact falls in the sequence matters more than who produced it — a property that makes the framework durable under changes in the identity of makers.

Sequences have shape. The early, middle, and late phases of a sequence have different formal properties, and the phase an artifact occupies determines what it can contribute to the sequence's development.

Multiple sequences operate simultaneously. A single period contains many overlapping sequences at different stages, and a single artifact may occupy positions in several sequences at once.

The framework survives AI. Because sequences are defined structurally rather than biographically, the arrival of non-human makers does not collapse the analytical apparatus — it extends it.

Debates & Critiques

The central contested application of the concept is whether AI systems participate in sequences in the same sense that human makers do. Critics argue that AI merely processes sequences without inhabiting them; defenders respond that the sequence's internal logic operates regardless of the maker's experience. Kubler's framework permits both readings but insists the distinction matters — not for whether AI can fill a sequence (it clearly can) but for whether it can perceive when a sequence is insufficient and open a new one.