The Logjam Metaphor for Concepts is perhaps too neat. The brain is even messier than that.
I like the logjam metaphor because it evokes the messiness of biological solutions to problems. In Associations of Dawn, I recounted how my logjam for the word "dawn" has, in the last ~40 years accumulated an association such that:
> When I read [a sentence like "The ship sails at dawn], almost invariably I’d get what I can only describe as a little flash of recognition that I was not just describing a time of day but also saying [Dawn's] name.
That's good, but it suggests there's a single *place*, the logjam, where a bunch of associations hang out together. It's a bit too much like a data structure with properties hanging off it, ready for lookup.
I doubt that’s right, because it doesn’t capture how *associative* and *distributed* the brain is.
Consider: some estimate that the average neuron is being fed impulses from 1000 other neurons. When I say “network” it’s easy to imagine a compact cluster of neurons. Some on the exterior are fed impulses from “upstream” neurons, others on the exterior feed impulses “downstream,” and the majority interior neurons are connected only to other cluster members.
Instead, your average neuron will be part of many networks, and networks overlap a good amount. There *are* large-scale structures in the brain, and neurons come in different types with different roles, and neurons aren’t the only cells with a role in neurotransmission and the formation of memory, but I want to emphasize that very many of the brain’s actions seem to require a whole bunch of networks activating, reinforcing, or dampening each other. William James describe a child’s first experience of the world as a “blooming, buzzing confusion,” and that also describes the brain that has to deal with it.
So a concept is less a place than **a pattern of activation** that lights up a bunch of neurons smeared through and overlapping with other concepts.