Recall performance is best if someone's state at the time of testing matches his state at the time of learning. A benefit of context reinstatement - that is, improved memory performance if we re-create the context that was in place during learning- can be demostrate. The context has its effect only because it influences how the person thinks about the materials to be remembered. Fisher and Craik presented their research participants with a series of word pairs. The participants were instructed to learn the second word in each pair. For half of the pairs, the "context word" was semantically assosicated with the target word. So they had to think about the word's meaning. For the other pairs, the context word was one that rhymed with the target. So they had to taking about the target word's sound. Conclusion was that there are two sepatate influences on memory working at the same time: an advantage for thinking about meaning and an advantage for matched learning and test conditions.