On filtering and abstracting

This is pure ignorant speculation, but I was interested in the linked article and its argument that much infant learning is so efficient (particularly, it suggests, early language learning) because the child's brain is insufficiently developed to filter information, and has to swallow it whole, as it were.

One issue that frequently comes up when covering the basics of memory as a precursor to learning theories is the existence (or not) and form of photographic (or eidetic) memory. I'm not going into detail because I do know I don't know enough about it even to be a reliable guide to other sites. Insofar as it is agreed to exist, it is much more common in children than in adults. That would seem to fit with the argument of the post and article. It appears that eidetic memory stands in the same relation to an "ordinary" memory as a picture of a document does to a version scanned with a print recognition utility. The eidetic memory will take up much more of (in this case a computer's) memory because it includes all the information of the document as a physical object (such as folds and smudges in the paper) regardless of their importance. The scanned text of course loses such irrelevant information, becomes much leaner, and of course editable.

I am reminded of Bruner on the enactive, iconic and symbolic modes of representation and the issue of whether being almost confined to an enactive mode (as an eidetic image is--granted that an image is by definition iconic to some degree) is as much a limitation to learning as a resource. Certainly that is the case for people with autistic spectrum disorders.