Research

P/A Institute develops philosophical, analytical, and computational methods for studying systems in which space, technology, artificial intelligence, civilization, and natural processes interact through heterogeneous relationships. The work proceeds through four connected research programs.

The shared research principle across all four programs is to observe, model, and audit the relational configurations that constitute space, technology, artificial intelligence, civilization, and natural processes — without fixing the analytical center, object boundaries, or method in advance. Models themselves are treated as provisional and subject to continuous audit.

Total Relational Configuration Analysis

How do heterogeneous actors — humans, AI systems, machines, institutions, capital, energy flows, logistic systems, and natural processes — form, maintain, and dissolve operational configurations without any single element serving as a fixed analytical center?

Existing analytical frameworks tend to fix a privileged analytical unit (the human subject, the firm, the state, the AI system) and read other elements as background or context. This misrepresents configurations in which the “central” element may itself be a dependent variable of surrounding processes.

Hyperflat relational mapping — treating all elements as variables with asymmetric relations (dependency, maintenance burden, control, energy supply, reproductive capacity). Configuration boundaries are treated as variable and revisable. No ontological priority assigned in advance.

Configuration · Variable Boundary · Hyperflat Analysis · Asymmetric Dependency · Maintenance and Stoppage · Reproductive Closure

Machine-Ecological Capital Theory, Hyperflat Distributed Mutual Audit framework, Scored-Agent analysis

Mapping a specific AI deployment as a full relational configuration: compute infrastructure, energy sources, human maintenance, institutional ownership, downstream scored populations, and feedback loops

Spatial Computing / Spatial UX

Core Project: When Space Becomes the Device

Under what technical, material, and computational conditions can physical space function as a complete interface — replacing or supplementing the smartphone and fixed display as the primary site of computation, display, input, and communication?

Information environments are currently enclosed within smartphone screens and fixed displays. This creates spatial, bodily, and attentional constraints. The question is not whether spatial computing is conceptually appealing, but what it would actually require: hardware specifications, sensing resolution, latency constraints, projection or display technologies, body-tracking accuracy, and AI processing requirements.

Systems analysis of current spatial computing technologies (AR/VR headsets, computer vision, LiDAR, spatial audio, projective display, environmental sensing). Gap analysis between current capabilities and full spatial interface implementation. Prototype design and feasibility mapping.

• Spatial smartphone / spatial tablet (freestanding spatial equivalents)
• Spatial UI and body-operated interaction
• Environmental computation and ambient display
• Projection-based interfaces
• AI-driven spatial recognition and response
• Sensor networks for environmental input

Conceptual system architecture for a spatial interface operating without a handheld device. Hardware requirement mapping. Identification of current technology gaps and plausible near-term implementations.

System design documentation, hardware gap analysis, interaction model prototypes

Specification of minimum viable spatial interface — defining the exact sensor, display, processing, and AI requirements for a functional spatial smartphone prototype

Total Relational Technosphere Archaeology

What can be reconstructed about the operational configurations of a civilization from its residual material traces, after civilizational disruption, human extinction, and geological transformation?

Standard archaeology reconstructs past human activity from physical traces. This research extends the problem: what happens when the civilization to be reconstructed operated through complex machine systems, AI, distributed computation, and industrial infrastructure — systems that may continue operating, degrading, or transforming in the absence of human maintenance?

Two complementary directions operated as one research program:

Forward Simulation: Model the sequential transformation of current technospheric configurations through disruption scenarios (localized nuclear war, systemic infrastructure failure) → civilizational reconfiguration → machine operation without human maintenance → operational extinction → residual technosphere → geophysical and astronomical transformation across deep time.

Inverse Reconstruction: Given a set of residual observations (physical traces, electromagnetic signatures, isotopic records, structural remains), reconstruct the lost operational configuration that produced them.

Milankovitch cycles, sea level change, glacial dynamics, geomagnetic variation, erosion, sedimentation, and tectonic processes are incorporated as scientifically constrained boundary conditions that transform residual configurations over time — not as deterministic triggers of civilizational collapse.

Theoretical framework for residual operational configuration analysis, deep-time modeling approach, nuclear winter scenario modeling

Simulation of a specific infrastructure system (power grid, data center network) under progressive maintenance withdrawal — modeling degradation timeline, residual operational states, and observable traces

AI, Scientific and Model Audit

How can AI systems, scoring mechanisms, scientific models, physical simulations, observational frameworks, and state spaces be treated as provisional, revisable objects of analysis — rather than final arbiters of truth — while maintaining maximum fidelity to empirical constraint?

AI evaluation systems frequently treat model outputs as ground truth rather than as provisional measurements subject to observational limits and conceptual presuppositions. Scientific models similarly risk being treated as transparent windows onto reality rather than as constructed tools with specific validity conditions, boundary assumptions, and potential failure modes.

Systematic audit of model assumptions, observational boundaries, concept formation, state-space definitions, and validity conditions. Distinguishes between: the training objective, the internal computational mechanism, the behavioral output, and the claimed epistemic status. Applies this distinction across AI systems, physical models, and simulation frameworks.

• AI evaluation and scoring systems
• Scientific models and physical simulations
• Observational boundaries and measurement frameworks
• Concept formation and classification systems
• State space definitions
• Validity conditions and failure modes

Scientific constraints are retained maximally. The audit does not retreat into systematic metaphysics that produces negligible empirical differences. It treats models, observations, and conceptual systems as provisional and revisable — not as inaccessible to criticism, and not as arbitrary constructions.

Dialectical Direction Audit Theory, Freedom-Justification rate analysis, Transcendental Reflexion System, Score-Subject analysis

Audit of a specific AI benchmark — examining what operational configuration the benchmark actually measures, what it excludes, and under what conditions the benchmark scores become misleading as indicators of the claimed capability

All published papers, working papers, and technical documents are available in the → Publications archive.