Cognitive Psychology
About

Recall vs. Recognition

Two fundamental forms of memory retrieval: recall requires generating information from memory, while recognition requires identifying previously encountered items from among alternatives.

Recall and recognition are the two major forms of memory retrieval, and they differ in both their cognitive demands and their underlying mechanisms. Recall requires generating a memory without the target being present (e.g., "What is the capital of France?"). Recognition requires identifying a previously encountered item when it is presented again (e.g., "Was Paris on the list you studied?"). Recognition is typically easier than recall, but understanding why requires examining the processes underlying each.

Generate-Recognize Theory

The classic generate-recognize theory proposes that recall involves two stages: generating candidate responses from memory and then recognizing the correct one among the candidates. Recognition requires only the second stage. This explains why recognition is typically easier — it eliminates the generation requirement. However, the theory cannot explain all findings, such as the recognition failure of recallable words (Tulving and Thomson, 1973), where words that can be recalled sometimes fail to be recognized.

Dual-Process Models of Recognition

Modern recognition memory theory distinguishes two processes: recollection (consciously remembering the encoding episode, including contextual details) and familiarity (a feeling of knowing without specific recollective detail). Recollection is a slow, all-or-nothing process that depends on the hippocampus. Familiarity is a fast, graded signal that depends on perirhinal cortex. Most recognition decisions can be based on either process, but the remember/know paradigm allows researchers to separately assess their contributions.

Signal Detection Model of Recognition d' = z(Hit Rate) − z(False Alarm Rate)

Recollection + Familiarity contribute to recognition:
P(old | old) = R + (1 − R) × F   [dual-process model]
Encoding Specificity and the Recall-Recognition Relationship

Tulving's encoding specificity principle predicts that recall can sometimes exceed recognition when recall cues reinstate encoding context better than the recognition test items do. The recognition failure of recallable words demonstrates this: a word encoded in a strong associative context (e.g., "train — BLACK") may be recalled given the cue "train" but not recognized when presented alone as "BLACK" because the recognition test provides a different context than encoding. This finding challenged simple generate-recognize models and highlighted the importance of encoding-retrieval match.

Applications

The recall-recognition distinction has important applied implications. Multiple-choice exams primarily test recognition, while essay exams require recall — they are not equivalent assessments of knowledge. Eyewitness identification procedures (lineup identification is recognition; free recall of events is recall) differ in their vulnerability to different types of error. Understanding the different retrieval demands helps design more effective testing, assessment, and investigative procedures.

Related Topics

External Links