The encoding specificity principle, proposed by Endel Tulving and Donald Thomson (1973), states that successful retrieval depends on the overlap between the information encoded at study and the information available at test. A retrieval cue is effective to the extent that it was encoded as part of the original memory trace. This principle shifted the focus of memory research from storage to the interaction between encoding and retrieval.
The Principle
Tulving formalized the insight that what is stored in memory is not the item alone but the item as it was encoded in its context. A retrieval cue is effective only if it provides access to the specific encoding of the target. This explains why the same cue can be effective or ineffective depending on whether it matches the original encoding context, and why different cues can be differentially effective for the same memory.
Context-Dependent Memory
Environmental context effects provide striking demonstrations. Godden and Baddeley (1975) had scuba divers learn word lists either underwater or on land and tested recall in the same or different environment. Recall was 30-40% better when the learning and testing environments matched. Similar effects have been demonstrated for room context, background music, and even ambient odor — though the effects are typically stronger for recall than recognition.
Internal physiological states can also serve as encoding context. State-dependent memory has been demonstrated for pharmacological states (information learned under the influence of alcohol or marijuana is better recalled in the same state), mood states (mood-congruent memory: sad moods facilitate recall of sad events), and even body position. These findings extend encoding specificity from external context to internal states, demonstrating that the "context" encoded with a memory is broadly defined.
Transfer-Appropriate Processing
Morris, Bransford, and Franks (1977) extended encoding specificity to processing operations through the concept of transfer-appropriate processing. Memory is best when the cognitive processes engaged at retrieval match those engaged at encoding. Semantic encoding is best for semantic tests, but phonological encoding is best for phonological tests (such as recognizing words that rhyme with studied words). This finding challenged the levels of processing framework's claim that deeper processing always produces better memory.
Implications for Education and Practice
Encoding specificity has practical implications for study strategies. Studying in conditions similar to testing conditions should improve performance. Generating multiple encoding contexts (studying in different environments, using different strategies) should create more retrieval routes and reduce context dependence. The testing effect — the finding that retrieval practice is more effective than restudying — may partly reflect the benefit of practicing retrieval in conditions similar to those of the final test.