A neurobiological interpretation of semiotics: meaning vs. representation

The author contends that there are no representations in brains, only meanings. A meaning is the focus of an activity pattern that occupies the entire available brain. It is constructed through intentional action and learning from the consequences of the action. Communication between brains requires the construction of representations that express meanings and elicit meanings in other brains. These representations have no meanings in themselves. One aim of artificial intelligence is to build “meaning machines”, that can initiate actions in the context of broadly assigned goals and learn about their environments from the received results of their own actions. A prototypical device is described that is based in experimental nonlinear brain dynamics and implemented with differential equations. Studies of the neurodynamics of sensory cortices during conditioned responses of trained animals to learned stimuli have supported the development of a mathematical model of the nonlinear dynamic processes by which meaning is constructed. A second order linear ordinary differential equation with compression of output by an asymmetric static sigmoid function describes each node as a neural population. The nodes are coupled by weighted connections, and the set of equations is solved by numerical integration. The parameters are optimized to give aperiodic attractors simulating observed brain activity. The attractors are stabilized with biologically modeled additive noise