What the research says about remembering words, and what it does not say.
Vocabulary learning is one of the better-studied corners of cognitive psychology, which means most of the advice you meet online has a real experiment behind it somewhere. It also means a lot of that advice has been stretched well past what the experiment showed.
These articles go back to the primary work: Hermann Ebbinghaus measuring his own forgetting in 1885, Allan Paivio’s dual coding theory in 1971, Atkinson and Raugh’s keyword method in 1975, Piotr Woźniak’s SM-2 algorithm in the late 1980s, and the modern FSRS scheduler that now ships inside Anki.
Where a finding is solid, it is stated plainly. Where the popular version has drifted from the study, that is stated too. The aim is that you can tell the difference between a technique with decades of replication behind it and a technique that simply sounds scientific.
Krashen's i+1 hypothesis says the best input is comprehensible except for one new piece. Fill in the Blank now prefers real example sentences where the blank is your only unknown word — computed free, on-device, from your own review history.
Ninety years of memory research say learning similar words together makes both harder to learn. LingoBlend now keeps confusable words apart while they're new, then contrasts them once you're ready — plus a new Tricky Pairs screen.
LingoBlend's vocabulary graph has its first consumer: Smart Blend now spends every translated word on vocabulary you're actually learning — skipping mastered words, resurfacing words due for review, and favoring words connected to what you know.
LingoBlend's text blending was rebuilt from the ground up: LingoBlend's own selector now chooses which words to blend, and the AI only translates them in context. The result reads better, handles idioms properly, and finishes about twice as fast.
A tour of the vocabulary graph now running under LingoBlend: why the words were chosen from real usage rather than textbook order, the kinds of connections that link them, and how inflections like despierta resolve to despertar.
Your brain stores words as a connected network — the mental lexicon — not a flat list. Here is what the research says, and why vocabulary apps should be built the same way.
Dual coding language learning pairs each word with a vivid image so verbal and visual memory reinforce each other. Learn the theory and how to apply it.
The keyword method links a foreign word to a sound-alike native word plus a vivid image. Learn Atkinson & Raugh's 1975 memory trick, with examples.
Spaced repetition for language learning explained: the forgetting curve, SM-2 vs FSRS vs Anki, and the daily review intervals that actually work.
Join LingoBlend and turn any text into a personalized language lesson. Free to start, no credit card required.