Connectionist Systems vs. Classical Models in Cognitive Science

Defending the Theory of Mind (IT)

The Theory of Mind (IT) can be defended in two ways:

• Impossibility of the PMM: It is impossible to determine if two subjects are in identical mental states, as these states are private and unique.
• PMM Credibility: Even if the PMM is accepted, it can be argued that this principle applies to individuals of different natures, such as animals (e.g., a rooster) and humans (e.g., Vladimir). In any case, IT implementation would need to be developed specifically for each cluster.

Classical Processing vs. Connectionist Models

Classical processing models are based on the idea that mental states are computational states defined by symbol manipulation. These systems operate under the premise that the brain functions like a computer. Conversely, connectionist systems represent a different approach:

• Classical Systems: Simulate the mind by manipulating symbols specific to a machine.
• Connectionist Systems: Simulate mental states based on structural models inspired by actual brain architecture, moving away from pure symbol manipulation.

Neurocomputational Representation and Theory of Mind

The Neurocomputational Theory of Mind (NWT) utilizes connectionist models that mirror brain architecture. In these systems, nodes function similarly to neurons, linked together to transmit signals.

Node Activation Mechanics

A node possesses multiple entries through which it receives signals. If the sum of these signals exceeds a specific activation value (known as force), the node activates. For example, if a node has a force of 3 and receives three input signals with values of 1, 2, and 2, the node will activate because the total exceeds the required force.

Network Structure

Connectionist networks are organized into three types of layers:

• Input Layer: Interfaces with external data.
• Output Layer: Provides the final processed result.
• Hidden Layer: Internal connections that link the input and output layers.