Sometimes I blank on a test and then remember it perfectly afterward, which doesn't feel like the memory was gone.
Accepted answer
Your experience is the textbook case for **retrieval failure**: the memory is still stored, but you can't access it right then. Tip-of-the-tongue states are the clearest evidence — you know you know it, and often recall partial information like the first letter.
The strongest support is that a **cue** frequently restores access. If information were truly gone from storage, no cue would help. Related effects: context-dependent memory (recall improves in the environment where you learned) and state-dependent memory (recall improves in a matching internal state).
Encoding failure is different — information that never got attentively processed was never stored, so no cue will help.