Are Numerical Symbols Fundamental to Neural Computation?

Abstract: Neuroclassicism is the view that cognition is computation and that core mental processes, such as perception, memory, and reasoning are products of digital computations realized in neural tissue. Cognitive psychologist C. R. Gallistel uses this classical framework to argue that all cognitive information processing is based on symbolic operations performed over quantitative values (i.e. numbers) stored in the brain, much like a digital computer. Assuming this hypothesis, he investigates how the brain stores quantitative information (i.e. the numerical symbols involved in neural computation). He claims that it is more plausible that memories for numbers are stored within molecular mechanisms inside the neuron, rather than within specific patterns of cell connectivity (the substrate for memory storage assumed by the traditional Hebbian plastic synapse model). In this paper, I dissect and critique Gallistel’s argument, which I find to be undermined by the findings of contemporary neuroscience