Prime Object

In the AI Story

The concept requires careful handling because it resists the categories that make art-historical and technological analysis comfortable. The prime object is not a masterpiece — masterpieces typically occupy late positions in sequences, after the formal possibilities have been extensively explored. The prime object is not an origin in the Romantic sense — it does not emerge from sovereign creative will but from the recognition that an existing landscape of solutions is insufficient. It is not the best exemplar of a type but the first exemplar, often crude, frequently misunderstood at the moment of its production, sometimes recognized as significant only in retrospect when the sequence it opened has developed enough to reveal what it made possible.

Kubler identified two modes through which prime objects emerge. The first is combinatorial: knowledge from one domain encounters a problem in another, and the encounter generates a solution that neither domain could have produced alone. The second is what Kubler called pure invention, in which the maker creates 'solely by means of his own engagement with his milieu,' producing a solution 'experientially and theoretically untied to earlier thinking.' Both modes share a prerequisite: the capacity to perceive that a new problem exists, that existing sequences cannot address it, and that a genuinely new class of solutions is required. This perception of structural absence — the felt recognition that the formal landscape lacks what is needed — is what AI has not yet demonstrated.

The asymptotic quality of AI's relationship to the prime object is the central tension of the current moment. Generative models produce outputs of extraordinary fluency by interpolating within and across the formal sequences their training data represents. The outputs are often surprising, occasionally magnificent, and structurally replicas — variations within existing sequences. The prime object, by definition, does not belong to the distribution defined by what preceded it; it opens a new distribution. This is not a limitation of current models that future models will obviously overcome. It is a structural feature of statistical inference, which produces outputs belonging to the distribution defined by the training data.

The examples that test this distinction most severely come from domains where AI's combinatorial power is most impressive. In drug discovery, AI has identified molecular candidates no human researcher proposed; in mathematics, it has generated proofs that surprised experts; in materials science, it has proposed structures that worked when synthesized. Kubler's framework forces the question: are these prime objects or sophisticated replicas? A new drug operating by a known mechanism, however optimized, fills a sequence. A drug operating by a genuinely new mechanism opens one. Careful analysis of the reported AI contributions suggests most fall in the first category. The question remains open, and Kubler's vocabulary makes it precisely answerable: can AI produce an artifact that opens a sequence the existing landscape did not imply?

Origin

Kubler articulated the concept most fully in chapter 2 of The Shape of Time, drawing on his work with pre-Columbian artifacts where the absence of biographical information forced attention to structural position. The concept was refined in later essays and developed implicitly through decades of his teaching at Yale, where students were trained to ask of any artifact not who made it but where it fell in the chain of solutions to the problem it addressed.

Key Ideas

First, not best. The prime object is defined by its position as the first in a new sequence, not by the quality of its execution — a distinction that separates structural significance from aesthetic achievement.

Demonstration, not refinement. The prime object's value lies in demonstrating that a new class of solutions exists; subsequent replicas realize the potential the prime object opened.

Two modes of emergence. Prime objects arise either through the confluence of previously separate sequences (combinatorial) or through pure invention responding to a perceived structural absence.

Structural absence as prerequisite. Opening a sequence requires perceiving that existing sequences are insufficient — a perception that is not a computation but an experience of inhabiting a landscape and finding it inadequate.

The AI question made precise. The debate over AI creativity becomes tractable when posed in Kublerian terms: can AI produce an artifact that opens a sequence the existing landscape does not imply?

Debates & Critiques

The unresolved question is whether the production of prime objects is structurally restricted to embedded biological intelligence or whether it is contingent on cognitive capacities that AI systems may eventually develop. Kubler's original framework assumed the former without argument. Contemporary readers cannot. The evidence as of 2026 supports a cautious version of the original assumption — AI produces extraordinary replicas but no convincingly demonstrated prime objects — but the history of claims about machine limitation counsels humility.