Are literal memory and high abstraction compatible ?

Page 1 of 1 [ 1 post ] 

Metaphile
Emu Egg
Emu Egg

User avatar

Joined: 1 May 2014
Gender: Male
Posts: 1

08 May 2014, 11:31 am

Hello,

I’m a NT French PhD student in cognitive psychology. I am working on the idea that we have two independent memory mechanisms, which I call association and abstraction.

- Association underlies episodic memory : the different perceptions are associated in a memory trace, and thinking of one cue will enable to recall the associated cues.

- Abstraction underlies the creation of new concepts by analogy (in a very broad sense, as defined in the last Hofstadter and Sander’s book Surfaces and Essences (2013) - I love this book, but I’m not very neutral, as Emmanuel Sander is my PhD director). An example of how a concept is built with the concept of mother : for some time in the life of an infant, Mother refers to a particular single person, as a proper noun. Then one day the child realizes that other children have a mother too : an abstract concept of mother is born, and refers to an adult woman who takes care of a child. The concept is then further extended when it is realized that every human, even adults, have a mother. But the abstraction doesn’t stop there at the notion of "female biological parent/female nurturer". With time, the term is further extended so that it encompasses terms like motherland, Mother Earth, "Greece is the mother of democracy", mother cell, or a mother board in a computer.
Viewed as a memory encoding mechanism, abstraction underlies semantic memory, but also recognition/categorization. To me, association occurs when one recognizes a situation as new : that would explain why we better remember the "first times". For example, we might be able to episodically remember the first time we ate Japanese food, but probably not the 7th time. When we recognize a situation as similar to one lived before, we will use abstraction to encode information and enrich our concept of that situation.

I know my choice of names is not great. Abstraction is very hard to accurately define (after all, it’s an abstract notion :wink: ), and the word association is generally associated with the idea of rewards and punishment, which are not in the scope of my theory.

The most striking case for this opposition is the case of Solomon Shereshevsky, the mnemonist of the famous book by Alexandre Luria. His literal memory seemed limitless : he was for example able to recall a list of random numbers learnt 15 years before (without rehearsing them). He was massively synesthete : numbers are transformed into shapes, tastes, sounds, etc. I interpret it like this : since everything is so distinctive, he always use his association mechanism, and encodes literally. The downside of his extreme memory is that he had abstraction difficulties : he had almost no understanding of figurative language and of abstract concepts he couldn’t visualise. He even had difficulties recognizing faces, finding them "constantly changing".
Another case which is coherent with my dichotomy is the Savant Kim Peek : almost perfect literal recall of thousands of books, but he seemed to lack understanding of metaphors and abstractions in general.

Autism, and probably in a lesser extent Asperger’s, seem to be associated with both
1) literal thinking (difficulties with understanding figurative language and with perceiving relevant links between different situations that doesn’t share the same context)
2) literal memory (accurate memory of specific details)
I have read some great discussions about these two traits on wrongplanet, but I haven’t found a conjoint discussion of the two. Do you think that those two traits generally co-occur ?

According to my theory, literal memory and high levels of abstract thinking (whatever that means - but basically, understanding relevant analogies) should be negatively correlated. Does that seem coherent with your personal experience ?

Thank you !