Final answer:
Sounds and graphemes are processed and committed to long-term memory through a three-stage encoding process. Semantic processing has been shown to be effective for long-term memory retention.
Step-by-step explanation:
Sounds and graphemes must be processed to commit a word to long-term memory. This is part of the three-stage process of memory encoding: sensory memory, short-term (also known as working) memory, and then long-term memory.
As explained in the Atkinson-Shiffrin model of human memory, information needs to be encoded, stored, and later retrieved. The initial processing usually happens automatically, but committing information to long-term memory often requires more effortful processing.
Studies by psychologists such as Craik and Tulving in their levels-of-processing framework have shown that semantic processing, referring to the meaning of words, can be particularly effective for long-term memory encoding, compared to acoustic or visual processing.
The long-term memory has a virtually unlimited storage capacity, in contrast to short-term memory, where information can decay within 18 seconds or be impacted by proactive interference as per the findings of Peterson and Peterson (1959), and Keppel and Underwood (1962).