Why intuitive theories of memory lead us wrong: memory representations are continuous strength, population-based and hierarchical
In the real world, objects are discrete physical entities – your coffee mug either is or is not in your hand. As a result, both in everyday life and in memory research, there is a tendency to use a physical metaphor to understand memory: people tend to think of an object they are trying to remember as either in mind or not in their mind, and to say that we hold items in mind, as we hold real objects in our hand. This metaphor serves as a core mental model used in most conceptions of memory: all-or-none, discrete, and functioning at the level of entire objects or other discrete representations, like chunks.
In this talk, I’ll argue for a new way of thinking about memory that strongly contrasts with this common and intuitive view. I’ll show that individuated items are far from the only kind of representation people form, and that it is necessary to consider interactions among an entire hierarchy of representations to understand memory even for a single item. I’ll also show that memory representations, even for single items, are population-based and continuous in strength. Altogether, I’ll argue that even for those interested in cognition, analogies from neuroscience — with population codes, hierarchical representations and noisy signals — best allow us to understand memory limits, rather than physical analogies about discrete items.