From engrams to schemas

One of the principal concepts instrumental in cognitive linguistic research is that of schemas, defined as general, partly abstract mental outlines shared by the memories of specific concepts and experiences. They are mental frameworks guiding the way we perceive and remember information, and they play a major role in language learning and processing. While the concept of schemas has been addressed by many authors, who have described their nature and functioning, our understanding of schematic knowledge is still relatively speculative. For example, to account for how schemas emerge, “lossy compression” is postulated, which means that abstraction is achieved through forgetting. Such conjectures can be tested against some recent insights from memory research on engrams, defined roughly as networks of neurons activated in the processing of specific pieces of information. This overview paper presents findings from engram research suggesting that abstract schemas form in ways other than simple forgetting or erasure of information from memory. Instead, abstraction is achieved by accumulating different representation strengths for memory traces underlying frequent and infrequent (incidental) information.