So far, we've learned about types of explicit memory. Memories for events, people, places, and objects are all types of explicit memory. Generally, an explicit memory is any memory accompanied by the experience of remembering. Memories of particular events are called episodic memories, as in, "I remember when my uncle took me to see the Red Sox." Memories for knowledge that you know you have but may not remember learning are called semantic memories, as in, "I remember that the first president of the United States was George Washington." Vocabulary is also a type of semantic memory, as in, "I remember that this object is called a chair."

However, not all memories are explicit. There is another system of memory that contains memories we don't even know that we have. Yet, these hidden memories can affect the actions we take and the choices we make. This system, only recently discovered, is called implicit memory. Implicit memory can cause priming, in which memories of specific items influence the way that people process information, without any conscious awareness of the effect. Implicit memory is also involved in procedural learning (learning a procedure such as how to ride a bike or how to program a computer).

One commonly used task that shows the priming effect of implicit memory is the stem-completion task. Subjects study a list of words (e.g. show, mock, cape). They are then given a standard (explicit) recognition memory test, in which they must choose all the words that appeared on the first list (e.g. Did you see the word "show"? Did you see the word "iron"?). Then, subjects are given an implicit memory test. In the stem-completion test, they are shown a list of incomplete words, called word stems (e.g. sh_ _, mo_ _, ca_ _) and are told to fill in the blanks with whatever word comes into their mind. Subjects fill in the blanks with words they studied more often than would be expected if they were filling them in randomly (e.g. they are more likely to complete the stem sh_ _ with "shoe than shot). This is true even when subjects did not recognize the words on the earlier explicit recognition test. In fact, subjects' performance on the two tests is completely unrelated; even though a subject may show no explicit memory for a word (e.g. doesn't recognize "shoe"), he may still show implicit memory (e.g. fill in sh_ _ with shoe) and vice versa. This dichotomy also shows up in other measures of implicit memory, suggesting that there are two separate systems responsible for processing explicit and implicit memory.

There is some biological evidence that also indicates a separate system for implicit memory. In amnesia, damage to parts of the brain (usually in the hippocampus and surrounding areas) can cause severe problems with explicit memory. In anterograde amnesia, people have trouble forming new explicit memories. For example, an amnesic person probably couldn't remember what they had for dinner last night (episodic) or the name of Russia's new president (semantic). However, even the amnesiacs with the most severe damage seem to have intact implicit memory. They perform the same as normal subjects on priming tasks such as the stem-completion, showing that their implicit memory system works the same way as other people's. Also, they are able to learn complex procedural tasks despite having no memory of the previous lesson. One researcher was even able to teach amnesiac patients to program a microcomputer. This means that the patients' procedural memories, part of the implicit memory system, must be working. People with anatomical damage that destroys their explicit memory nevertheless show intact implicit memory. Thus, the two systems, implicit memory and explicit memory, must be separated in the brain.