The Agential Cut

In the AI Story

In quantum mechanics, Bohr demonstrated that the properties of quantum objects — position, momentum, spin — become determinate only through specific experimental apparatuses. An electron does not possess a definite position until a position-measuring apparatus is employed, and the apparatus that measures position is materially incompatible with the apparatus that measures momentum. The boundary between the object being measured and the instrument measuring it is not a pre-given fact of nature but is enacted through the specific material arrangement of the experiment. Barad's agential cut names this enactment, generalizing it beyond quantum physics to every domain where boundaries come into being.

Applied to the practice of authorship in the AI age, the concept reveals that every act of attributing a book, a piece of code, or an idea to an identifiable author is an agential cut. The human and the machine who collaborated on the work did not exist as determinate, separable contributors during the unfolding of the collaboration — their contributions were entangled, mutually shaping. The name on the cover performs a boundary-making practice after the fact, producing the author as a determinate entity with determinate responsibility. The cut is necessary (someone must be accountable) but it conceals the entanglement that produced the work.

The ethical weight of the cut is central to Barad's framework. Every cut includes something and excludes something, and the exclusions carry consequences. The cut that assigns all creative credit to the human author excludes the machine's contributions from recognition — and also excludes the machine's designers from responsibility when the output fails. The cut that naturalizes the boundary between backend engineer and frontend developer in the pre-AI workplace produced a particular professional identity, closed off certain possibilities, and rendered others invisible. When the apparatus changed — when Claude Code dissolved the translation costs between domains — the cut was re-enacted differently, producing a different set of professional subjects.

The framework has direct practical implications for AI governance. If the boundaries AI systems produce (between relevant and irrelevant, between deserving and undeserving, between signal and noise) are agential cuts rather than neutral determinations, then the design of AI systems is the design of boundary-making practices with ethical consequences that cannot be discharged through claims of algorithmic neutrality. Dan McQuillan has pressed this point: setting up an AI system one way or another changes what becomes naturalized and what becomes problematized, and who gets to set up the AI becomes a question of power.

Origin

The concept was introduced in Barad's 1996 essay 'Meeting the Universe Halfway' and developed systematically in the 2007 book of the same title. It draws on Bohr's principle of complementarity but transforms it: where Bohr treated the cut as a pragmatic feature of experimental practice, Barad argues that cuts are ontologically constitutive — producing determinate entities from entangled phenomena rather than merely distinguishing pre-existing ones.

Key Ideas

Cuts produce distinctions. Before the cut, what becomes two entities is entangled; after the cut, each is determinate in its own properties.

Cuts are real. They have material consequences — different cuts produce different measurements, different subjects, different worlds.

Cuts are not natural. They are practices enacted through specific apparatuses, and different apparatuses enact different cuts.

Cuts carry ethical weight. What a cut includes and excludes is a matter of responsibility — the cut could have been enacted differently, and that counterfactual carries moral force.

Cuts must be maintained. The boundary between understanding and generation, between chosen continuation and compulsive extension, does not persist on its own — it requires ongoing practice to preserve.