°°°°~x§x-<@>
Abstract: This article discuss the epistemology of modeling and simulations, positioning them not merely as descriptive tools, but as crucial heuristic devices and active epistemological practices central to developing a comprehensive, enacted approach to cognition and self-understanding. This focus is situated within the larger framework of Second-Order Rationality² and the integration of first-person (1P) and third-person (3P) perspective.
N.B.: All these considerations on my blog are not a priori theoretical research but post hoc forms slowly implemented and stabilized, coming from my own life experience and computer experimentations with these kind of models.
1. Models and Simulations as Epistemological Tools
In the general context of scientific thinking, models, modeling, and technical metaphors play a crucial heuristic role. From the perspective of scybernethics, theories, paradigms, and models are viewed as “attentional grids”—ways of seeing and thinking that shape perception and interpretation of phenomena.
A primary focus is advocating for an epistemology of modeling and simulations of cognitive artifacts. This urgent necessity arises from the need to stop the growing confusion between models and empirical reality. The traditional scientific commitment implies that science does not describe the “world-by-itself” but, more subtly, our rational interaction, technologically mediated, with the world
The purpose of modeling must shift from solely explaining (E-models) toward understanding (U-models), which has historically been in the blind spot of traditional scientific modeling. U-models are comprehensive enough to integrate and include E-models.
2. The Role of Experimental and Experiential Epistemology
The use of computer models and simulations is articulated through the framework of “experimental epistemology” (as inspired by McCulloch) and “experiential epistemology”. This perspective emphasizes that the decisive thing with modeling is not the model per se, but what the model and working with the model does to our mind.
This approach is captured by the idea of moving “From computers which think to computers which make me think” (Cf. Bachimont).
1. Experimental Epistemology (3P focus): Designing cognitive complex models and simulations allows the observer-actor to see the global bottom-up/emerging and behavioral consequences of local functional hypotheses, which might otherwise be impossible to deduce intellectually or mathematically.
2. Experiential Epistemology (1P focus): Working with simulations, particularly those based on quasi-analogic paradigms like Parallel Distributed Processing (PDP) connectionist models, leads to a trial/error converging experiential epistemology. By consciously using the illusory analogical power of these simulations, the observer-actor can reflect on their own thinking processes by feeling “gestural”/sensori-motor schemes differentiations. This allows for the self-understanding and objectification of the mechanical dimension of one’s own thinking.
3. The Epistemological Significance of PDP / Connectionist Models
The Scybernethics approach highlight Parallel Distributed Processing (PDP) models (also known as Artificial Neural Networks or “deep/machine learning”) as particularly valuable epistemological tools because they can be interpreted in a way that resonates with enaction and phenomenology.
Instead of viewing them merely through a statistical or associationist lens, PDP models are interpreted as dynamic systems that:
• Simulate Analogical Bioprocessing: They simulate the scientifically repressed and intellectually unthinkable analogical bioprocessing of living cognition by processing in a parallel and distributed manner.
• Model Phenomenological Intentionality: They can be interpreted as models of enacted phenomenological intentionality, serving as an elegant computational model for the enacted emergence of natural categories as prototypes (Rosch).
• Offer Post-Functionalism: They allow for a post-functionalist conception where cognitive functions and “inferences” are emergent and enacted by the network’s learning activity, suggesting that our own logical reasoning is a posteriori of sensorimotor activities
Because the intricate dynamics of these systems cannot be fully understood intellectually or linearly, only the combined practice of design, observation, and epistemological culture allows for an intuitive and heuristic approach to access and make intelligible their ana-logical enactions.
4. Modeling within the Epistemological Approach (1P-3P Circulation)
This specific focus on modeling is integrated into the holistic Scybernethics (Second-Order Rationality²) approach, which relies on balancing dualities and integrating perspectives.
Modeling becomes a key driver of the hermeneutic/co-determining circulation between the two main epistemological poles:
1. First-Person (1P) Perspective: The embodied or phenomenological point of view, concerning the lived experience and self-understanding.
2. Third-Person (3P) Perspective: The abstract objective science/Cartesian/Cognitivist perspective, concerning formal, intersubjective knowledge, including scientific culture and epistemology.
In this framework, modeling becomes an active epistemological practice driven toward self-reflection. The iterative reflection on the model and the model of the modeling points recursively toward self-understanding. This is part of the “ambijective gesture,” a self-conscious cyclical process between objectification and phenomenological subjectification that co-determines the epistemological context and body-based meaning-making.
By embracing this 1P-3P methodology, the observer-actor explicitly includes themselves in the system being observed, aligning with the principle of second-order cybernetics. This realization fosters epistemological flexibility, enabling the individual to circulate effectively between complementary modeling styles and relate them to their own understanding.
5. Critique of Reductionist and Confusing Modeling Styles
The centrality of modeling epistemology also informs critiques of other popular approaches:
• Critique of PP-FEP: Predictive Processing/Free Energy Principle (PP-FEP) modeling is viewed critically. While operationally effective for communication (“convincing”/selling), these statistical constructs are considered misleading and poorly ethical/rational if they are interpreted as solely descriptive of underlying biological processes. The normative mathematical description of FEP is often just an a posteriori rationalization of self-organizational processes.
• The Technology-by-Itself Fallacy: A crucial epistemological concern is the “technology-by-itself” view subterfuge, where technology is mistakenly seen as neutral and self-operating. This leads to the illusion of interpreting complex cognitive simulations (like AI) as if they were independent entities, obscuring the human interaction and specific intentions embedded in the design and interpretation of the model.
6. Conclusion
Ultimately, for the enactive paradigm to succeed scientifically, it must integrate these modeling practices in a reflective way, avoiding the pitfalls of reductionism and the confusion between simulation and reality, while continuously coupling disciplinary knowledge with the self-reflexive understanding of the observer-actor.
°°°°~x§x-<@>