Algorithmic Objects

Ontological claim

Algorithmic objects are Luciana Parisi's answer to accounts that treat algorithms either as finite instructions representing material processes or as software dissolved into an all-encompassing material flow. Algorithms are data actualities: sequential spatiotemporal structures with effects, histories and powers of selection that are not reducible to a biophysical referent or to a timeless mathematical form (Luciana Parisi/Texts/Books/Author/Luciana Parisi - Contagious Architecture_ Computation, Aesthetics, and Space.pdf, pp. 21–23).

The claim is neither that code is alive nor that software is conscious. In The Holes in the Machine, Parisi contrasts instructions waiting to be performed with algorithms understood as information objects that select, evaluate and produce data. Her deliberately blunt formulation—algorithms “are things that think”—attributes an operative, object-like actuality to them without granting them a human subject or an integrated mind (Luciana Parisi/Texts/Interviews/The Holes in the Machine.pdf, pp. 1–4).

Prehension and incomputability

Parisi redeploys Alfred North Whitehead's prehension to describe this operation. An algorithm takes account of past data, present inputs and abstract possibilities; that selection irreversibly changes the actuality doing the selecting. Algorithmic objects can therefore be finite actualities infected with the incomputable data of eternal objects, their bounded procedures containing quantities that cannot be compressed into a smaller rule (Luciana Parisi/Texts/Books/Author/Luciana Parisi - Contagious Architecture_ Computation, Aesthetics, and Space.pdf, pp. 82–83).

The incomputable is internal rather than a wall where computation stops. Gödelian undecidability, Turing's halting problem and Gregory Chaitin's algorithmic randomness show that finite axiom systems contain irreducible quantities. Parisi uses that result ontologically: algorithmic parts can be larger than the wholes built from them, forcing procedures to encounter patternless data and revise their operations (Luciana Parisi/Texts/Books/Author/Luciana Parisi - Contagious Architecture_ Computation, Aesthetics, and Space.pdf, pp. 60–63).

Architecture and governance

Algorithmic architecture makes these objects spatially consequential. Parametric and interactive systems are often criticized as a smooth topology of neoliberal control, but Parisi argues that such critique misses the internal inconsistency of the computational apparatus. Algorithmic objects invisibly structure sociality while random or incompressible quantities fracture the system's claim to totality (Luciana Parisi/Texts/Essays/Algorithmic Architecture.pdf, pp. 2–4).

This is why Parisi calls patternless data the “contagious architectures of the present.” Algorithms persist as objective data even after a particular execution ends: later procedures inherit, evaluate and select them, adding new spatiotemporalities to computational culture (Luciana Parisi/Texts/Essays/Algorithmic Architecture.pdf, pp. 3–4). Contagious architecture names this infection of formal procedure, while soft thought names the noncognitive mode of decision that results.

CONTRADICTION Algorithmic objects are implicated in computational governance and can automate selection at infrastructural scale, yet their incomputable parts prevent them from forming the seamless predictive whole attributed to algorithmic control (Luciana Parisi/Texts/Essays/Algorithmic Architecture.pdf, pp. 2–4). Their partial autonomy is neither political innocence nor total programmability.