A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
High-frequency words are recognized, named, and retrieved faster than low-frequency words, reflecting stronger activation of frequently-encountered lexical representations. Frequency effects appear early in processing and persist across tasks. This demonstrates lexical access is probabilistic and shaped by experience statistics.
From your study of language comprehension, you know that understanding spoken or written language involves rapidly mapping sounds or letter strings onto meaning. The mental lexicon — the brain's stored inventory of word forms, their pronunciations, and their meanings — is the storehouse this process draws on. Lexical frequency effects reveal something fundamental about how that storehouse is organized: it is not a flat dictionary where every word is equally accessible. Instead, words are retrieved at speeds that reflect their frequency of encounter in a speaker's lifetime.
The evidence comes most cleanly from the lexical decision task: participants see a string of letters on a screen and press one button if it's a real word, another if it isn't. High-frequency words like "table" or "house" are recognized in roughly 500–600 milliseconds; low-frequency words like "flagon" or "effigy" take 100–200 ms longer. The same pattern appears in naming tasks (reading words aloud), priming paradigms, and even in production (how quickly you can retrieve and say a word in conversation). The effect is not just faster — frequency also reduces error rates and influences the earliest electrophysiological markers of word recognition (the N200 and N400 components in EEG), suggesting frequency shapes the recognition process itself rather than just a post-recognition decision.
The most influential account is activation threshold theory: each word in the lexicon has a resting activation level, and recognition occurs when that level crosses a threshold in response to bottom-up input. Frequent words have higher resting activation — their neural representations have been strengthened by repeated use, just as frequently-traveled neural pathways become more efficient. This is essentially a lexical analog of the synaptic strengthening principle: more frequent activation lowers the effective threshold for future activation. An alternative account emphasizes search models — the idea that the lexicon is searched in frequency order, so common words are found earlier. Modern connectionist models generally favor the activation account, as they naturally predict graded, continuous effects rather than discrete search steps.
Frequency effects are not purely a relic of past exposure — they are cumulative and ongoing. A word you've read 1,000 times is faster to recognize than one you've read 100 times, and learning a new word increases its recognition speed as you encounter it more. This has practical implications: vocabulary acquisition is partly a matter of encountering words enough times that their representations become robust. Reading volume in childhood predicts vocabulary breadth in large part because wide reading provides high-frequency exposure to a broad range of words. The frequency effect also helps explain why reading fluency develops over years — high-frequency words in a language become recognized at near-automatic speeds, freeing cognitive resources for comprehension rather than decoding.