The Keyword Method is a mnemonic especially suited to vocabulary learning that works by linking a new word to a familiar-sounding English word (the keyword) through a vivid mental image. It exploits acoustic similarity to create a memorable bridge between an unfamiliar form and its meaning. The procedure has three steps. First, the learner identifies a familiar English word that sounds like part of the new word—this is the keyword. Second, the learner looks up or recalls the translation. Third, the learner creates a vivid interactive image linking the keyword's meaning with the meaning of the translation. The Spanish mesa (table) resembles the English word "mesa" associated with a flat-topped geographical feature; for the more challenging Spanish cobarde (coward), the learner might picture a cob (horse) wearing armor but running away, integrating the acoustic and semantic components into a single bizarre scene. When no obvious phonetic match exists, the learner breaks the word into syllables or emphasizes part of it—the French fourchette ("fork") becomes "four-chette," visualized as a fork bearing the numeral 4.
The technique became one of the most studied in cognitive psychology through the work of Richard C. Atkinson and his colleagues in the 1970s. In the classic experiments by Raugh and Atkinson (1975, Spanish) and Atkinson and Raugh (1975, Russian), keyword learners substantially outperformed rote learners, with reported recall of roughly 70% versus 45% under some conditions. The advantage held across both immediate and delayed tests. Compared with pure semantic encoding, which ties new words to meaning through definition and context, the Keyword Method uses acoustic plus imaginal encoding, and many learners find it faster for the initial acquisition of large vocabulary sets. Studies show that keyword memories persist significantly longer than rote ones, especially after delays of days or weeks.
The Keyword Method extends well beyond foreign languages. It is widely used for scientific terms, medical terminology, brand names, people's names, and technical jargon, since any unfamiliar-form/meaning pair benefits from the same acoustic-imaginal bridge: "mitosis" linked to "my toes-is," "tungsten" visualized as a "tongue-sten" made of dense grey metal. The method works best for concrete nouns and short words, and finding good keywords for very short or unusual words can be difficult. For abstract concepts, the learner must construct creative concrete imagery—a bird escaping an open cage for "freedom," a balanced scale for "justice"—and use the same image consistently across encounters. For idioms and multi-word expressions, the most distinctive word supplies the keyword, with extended expressions decomposed into smaller memorable units. Noun gender can be encoded by adding a feminine or masculine figure to the keyword scene. Bizarreness is often recommended, but the experimental literature is more modest than the popular claim: the bizarreness effect shows up mainly in mixed lists where bizarre and ordinary items are studied together and is small or absent when a whole list is bizarre; what reliably helps is that the keyword and the meaning interact in one integrated scene. The method also supports pronunciation and tonal patterns when exaggerated visual mnemonics represent pitch contours. It can be combined with spaced repetition apps, which optimize review timing, while keywords optimize initial encoding. For children, simple, silly images featuring familiar characters often produce especially strong retention.