Why do connectionist models describe thought as patterns of activity across many units?

In the philosophy of mind and cognitive science, connectionist models—also known as parallel distributed processing—describe thought as patterns of activity across many units because they are designed to imitate the biological structure of the human brain.12

Unlike symbolic approaches that process information through step-by-step logical rules, connectionism posits that cognition emerges from the collective, simultaneous activity of simple processing units (artificial neurons) connected in a network. Key reasons for this description include:2

• Distributed Representation: Information is not stored in a single location but is spread across the entire pattern of connection weights within the network. This mirrors how the human brain appears to store information.3
• Parallel Processing: Computational tasks, such as facial recognition or language learning, are performed by generating firing activity across the network simultaneously rather than through a serial sequence of inferences.2
• Emergent Learning: Through training, the network adjusts the strength of connections between units based on experience. Thoughts—such as the concept that "dogs are furry"—are represented by these specific, simultaneous patterns of neural firing activity.2

By modeling the brain as a system of interconnected processors, connectionists argue that complex mental functions arise from the collective interaction of these units.23