Cognitive Science: Embodiment and Situatedness
This text is a preview of History of the Reception of Hegel, volume II, to be published in 2026 by Reflexivity Press.
The Critique of Cognitivism
In the 1980s, criticism formed against the cognitivist paradigm that had shaped early AI. Cognitivism says: the mind is a computer that processes symbols. The critics said: that is a poor model for human intelligence.
Hubert Dreyfus argued with Heidegger: human understanding is not symbol processing but being-in-the-world. We do not understand the world by manipulating representations, but by acting within it. A hammer is not understood through a description, but through use.
Francisco Varela (whom we have encountered as a theorist of autopoiesis) argued biologically: cognition is not computation but embodied action. The brain is not a computer that processes input and produces output. It is part of a living body that acts in an environment.
Embodied Cognition
The alternative to cognitivism is called “embodied cognition.” Its thesis: thinking is not detached from the body but shaped by it.
Experiments showed: the way we think depends on how we are embodied. When people furrow their brows (an expression of disapproval), they evaluate statements more negatively. When they nod, they are more likely to agree. The body thinks along.
This is anti-Cartesian: it is not the disembodied mind that thinks, but the embodied human being. And it is Hegelian: mind is not outside nature but its highest development. The body is not a prison of the mind but its medium.
The Consequence for AI
If thinking is embodied, then a disembodied machine cannot “truly” think — at least not as humans do. Symbolic AI misses something fundamental: embeddedness in world, body, situation.
From this arose new approaches: robotics that incorporates body and environment; situated AI that does not compute abstractly but acts concretely; neural networks that operate not symbolically but connectionistically.
These developments lead into Volume III. Here it suffices to note: the philosophical critique of early AI was justified. It did not prove that machines can never “think.” But it showed that thinking is more complex than cognitivism assumed.