Agential Realism

In the AI Story

Agential realism emerged from Barad's training as a theoretical physicist and her decades-long engagement with Bohr's philosophy-physics. Bohr had demonstrated that quantum properties are not inherent attributes waiting to be discovered but are produced through specific experimental apparatuses — an electron's position and momentum cannot be simultaneously determined because the material configurations required to measure them are mutually exclusive. Barad generalized this insight: if the apparatus co-constitutes the phenomenon in physics, then the boundary between observer and observed is a cut enacted through practice rather than a pre-given fact of nature. The implications extend far beyond the laboratory to every domain where knowing occurs.

The framework distinguishes sharply from representationalism, the dominant assumption in Western epistemology that the world consists of pre-existing entities waiting to be accurately represented. Barad's performative alternative recognizes that the world does not consist of entities awaiting representation but of phenomena produced through specific material-discursive practices. This is not relativism — the cuts are real and have material consequences — but it is a refusal of the assumption that there is a God's-eye view from which the world appears fully formed, awaiting neutral observation.

Applied to the AI transition documented in The Orange Pill, agential realism reveals that questions like 'Will AI replace human workers?' operate within a representationalist paradigm that assumes stable categories of human and machine. The performative alternative asks instead what phenomena are produced through specific configurations of human-machine entanglement, what those phenomena include and exclude, and who bears responsibility for the cuts that constitute them. The reframing transforms AI governance from a question of managing pre-given entities to one of attending to the apparatuses through which those entities come to be.

Scholars including Dan McQuillan, Inês Hipólito, and Eleanor Drage and Federica Frabetti have extended agential realism into the specific domain of AI systems, arguing that machine learning models do not merely represent the world but participate in producing it — and that the question of who configures the apparatus becomes the central political question of the age.

Origin

The framework took definitive form in Barad's 2007 landmark Meeting the Universe Halfway, though its elements had been developed across two decades of prior essays. The book synthesized Bohr's philosophy-physics, Michel Foucault's work on discourse, and Judith Butler's theory of performativity into a unified framework that could address phenomena ranging from ultrasound imaging to the measurement of quantum spin to the production of gender.

The framework's reception has been unusual in its breadth. It has become foundational across science and technology studies, feminist theory, new materialism, and posthumanist philosophy, and has been increasingly applied to questions of AI, algorithmic governance, and computational media — domains Barad herself did not initially address but for which her framework provides analytical tools of unusual precision.

Key Ideas

Phenomena, not objects. The basic units of reality are entangled configurations of matter and meaning, not independent entities with fixed properties.

Intra-action, not interaction. Entities do not pre-exist their relationships; they are constituted through them.

The apparatus is constitutive. Measuring instruments, cognitive frameworks, and technological systems participate in producing the phenomena they engage, rather than passively observing a pre-given reality.

Cuts are enacted, not discovered. The boundaries that separate one entity from another are practices performed through specific material-discursive configurations.

Ethics, ontology, and epistemology are entangled. How we know, what is, and what we value are co-constituted rather than independent domains.

Debates & Critiques

Critics including Caroline Braunmühl have argued that Barad's dissolution of pre-given subjects threatens the ethical and political vocabulary of suffering, rights, and recognition — that without a stable subject, the distinction between entities that can suffer and entities that cannot collapses. Barad's defenders respond that agential realism preserves the subject as constituted through entanglement rather than eliminating it, and that the reconception strengthens rather than weakens ethical response-ability.