Read More ↓
Recursive Simulations
The emergence of simulation as an epistemological technology, from scientific simulation to virtual environments
Perhaps the philosophy of simulation begins with the beginnings of philosophy itself; a foundational paranoia
Foundations of Western philosophy are based on a deep suspicion of simulations. In Plato’s allegorical cave, the differentiation of the world from its doubles, its form and its shadows, takes priority for the pursuit of knowledge. Today, however, the real comes to comprehend itself through its doubles: the simulation is the path toward knowledge, not away from it.
From Anthropology to Zoology, every discipline produces, models and validates knowledge through simulations. Simulations are technologies to think with, and in this sense they are fundamental epistemological technologies, and yet they are deeply under examined. They are a practice without a theory.
Some computational simulations are designed as immersive virtual environments where experience is artificialized. At the same time, scientific simulations do the opposite of creating deceptive illusions; they are the means by which otherwise inconceivable underlying realities are accessible to thought. From the infinitesimally small in the quantum realm to the inconceivably large in the astro-cosmological realm, computational simulations are not just a tool; they are a technology for knowing what is otherwise unthinkable.
Simulations do more than represent; they are also active and interactive. The “Recursive” in our thematic title, Recursive Simulations, refers to simulations that not only depict the world, but act back upon what they simulate, completing a cybernetic cycle of sensing and governing. They are more “an engine than a camera” as the saying goes. Recursive Simulations include everything from financial models to digital twins, user interfaces to prophetic stories. They cannot help but transform the thing they model, which in turn transforms the model and the modeled in an cyclical loop.
Recursion and Reflexivity
Recursion can be direct or indirect. It can be a literal sensing/actuation cycle, or the indirect negotiation of interpretation and response. The most nuanced recursions are reflexive. They mobilize action to fulfill or prevent a future that is implied by a simulation. Climate politics exemplifies the reflectivity of recursive simulations: through planetary computation, climate science produces simulations of near term planetary futures, the implications of which may be devastating. In turn, climate politics attempts to build planetary politics and planetary technologies in response to those implications and thereby extraordinary political agency is assigned to simulations. The future not only depends on them, it is defined by them.
Computational simulations as experiential, epistemological, scientific and political forms, but a framework is needed to understand these in relation to one another.
All of these forms of simulation pose fundamental questions about sensing and sensibility, world-knowing and worldmaking. They also have different relations to the real. While scientific simulations pose meaningful correspondence with the natural world and provide access to ground truths that would be otherwise inconceivable, virtual and augmented reality produce embodied experiences of simulated environments that purposefully take leave of ground truth.
These two forms of simulation have inverse epistemological implications: one makes an otherwise inaccessible reality perceivable, while the other bends reality to suit what one wants to see. In between is where we live.
Antikythera focuses on several emerging areas of Recursive Simulations research:
●
Scientific Simulations and physical modeling
●
Virtual and Augmented Reality
●
Simulation and Recursion Theory
●
Toy World Simulations and Digital Twins
●
Simulation and Theory of Mind
●
Archaeological and Anthropological Simulation
●
Geopolitics of Planetary Simulation Model
●
Simulation as Political and Logistical Technology
Friends from Neuroscience (and Artificial Intelligence) may raise the point that simulation is not only a kind of external technology with which intelligence figures out the world, but simulations are how minds have intelligence at all. The cortical columns of animal brains are constantly predicting what will be next, running through little simulations of the world and the immediate future, resolving them with new inputs and even competing with each other to organize perception and action.
For many computational simulations, their purpose is as a model that reflects reality (such as for climate science or astrophysics) For others the back and forth is not just mirroring. Some simulations not only model they world, but feedback upon what they model both directly and indirectly, they are recursive simulations which not only model an external reality but which directly act back upon that reality in a decisive feedback loop. What are called “digital twins” are one kind of such dynamic. In the recursive relation between simulation and the real, the real is the baseline model for simulations and simulations as a baseline model for the real.
Many AIs, especially those embodied in the world, such as driverless cars, are trained in Toy World simulations, where they can explore more freely, bumping into the walls, until they, like us, learn the best ways to perceive, model, and predict the real world.
Toy Worlds are where some AI’s learn to navigate the real world by navigating focused, reductive simulations of its contours. But these are not always closed. For AIs the boundary between a simulated world built of data and the real world perceivable as data is not always clear. For those training AIs to support physical actions in the physical world, this fuzziness can be leveraged. In simulated worlds, time can be speeded up, multiple generations and iterations can spawn in an instant.
Toy Worlds serve as a bounded domain of constrained information exchange and interaction between otherwise unlike and incompatible things and actions. The sim-to-real passage occurs not only in terms of the implications of specific learned expertise, but also through the virtual- physical hybridization of direct inputs and outputs: AI’s interacting with blends of both real and imaginary contexts and collaborators at the same time.
All of this includes what we recognize as robotics, but these embodied simulations are themselves being remade in the image of AI. Foundation Models become specialized, combining multiple modes of information and application. While machine learning already exists in the world, embedded across an artificial megastructure encrusting the planet, AI is on its way to become a generic solvent, soaked into things and how they behave. And so, the back and forth learning between artificial intelligence and natural intelligence never stops.
As our project Vivarium shows, we can consider multiple possible combinations: of human users, AIs as prostheses, AIs as users, human or humans as prostheses. This implies multiple combinations of embodiment, agency and action in and across the simulation, the real and the recursion.
The AIs world is a simulation of ours, but one we can interact with. For the AIs, our world is part of the omnisimulation that it calls simply reality.
Simulations as Epoch and Epistemology
We live in an era of highly politicized simulations, for good and ill. The role of climate simulations for planetary governance is only the tip of the proverbial iceberg. Antikythera considers computational simulations as experiential, epistemological, scientific and political forms and develops a framework to understand these in relation to one another.
The politics of simulation, more specifically, is based on recursion. This extends from political simulations to logistical simulations to financial simulations to experiential simulations: the model affects the modeled.
Antikythera’s research in this area draws on different forms of simulation and simulation technologies.These include machine sensing technologies (vision, sound, touch, etc.), synthetic experiences (including VR/AR), strategic scenario modeling (gaming, agent based systems), active simulations of complex architectures (digital twins), and computational simulations of natural systems enabling scientific inquiry & foresight (climate models and cellular/genomic simulations). All of these pose fundamental questions about sensing and sensibility, world-knowing and worldmaking.
They all have different relations to the Real. While scientific simulations pose meaningful correspondence with the natural world and provide access to ground truths that would be otherwise inconceivable, virtual and augmented reality produce embodied experiences of simulated environments that purposefully take leave of ground truth. These two forms of simulation have inverse epistemological implications: one makes an otherwise inaccessible reality perceivable, while the other bends reality to suit what one wants to see. In between is where we live.