About
Before you rely on this: these lessons were written with AI assistance. The sources cited in each one are real, but the surrounding explanation hasn't been reviewed by a linguist or German teacher. If something here matters to you, check the citation, not just the page.
The reasoning behind the method: why frequency-first, why interleaved, why spaced repetition.
What Zipf's law has to do with learning German
In 1935, the linguist George Zipf noticed something odd about word frequency in a language: if you rank every word by how often it’s used, the most common word shows up roughly twice as often as the second most common, three times as often as the third, and so on. Rank times frequency comes out close to a constant. That pattern, now called Zipf’s law, holds up across languages, including German.
| Rank | Word | Roughly how often, relative to #1 |
|---|---|---|
| 1 | die Zeit (“time”) | 1× |
| 2 | die Würde (“dignity”) | ~½× |
| 3 | das Leben (“life”) | ~⅓× |
| 4 | die Macht (“power”) | ~¼× |
| 5 | das Jahr (“year”) | ~⅕× |
The practical consequence is bigger than the trivia: a small set of high-frequency words accounts for a hugely disproportionate share of everything you’ll actually read or hear. Learn those first, and you’re covering a lot of real German quickly. Learn them alphabetically, or grouped by “beginner topics” like colors and animals, and that leverage is left on the table: sorted alphabetically, “die Ameise” (ant) and “die Zeit” (time) look equally important, even though one of them you’ll use constantly and the other almost never.
How this app uses it
Every noun in the curriculum carries a Zipf score, and new words unlock in that order, a batch at a time, not all at once. You’re never staring down 279 nouns on day one. You’re working through the ones that actually pull weight in real German, then the next tier, then the next.
Where the numbers come from
The word list and frequency scores in this app aren’t hand-picked. Each
noun’s zipf score comes from wordfreq,
an open-source word frequency database built by Robyn Speer from large
text corpora (subtitles, web text, books, and more) across dozens of
languages, including German.
wordfreq reports frequency on the Zipf scale: a logarithmic scale
where a word’s score is log10(occurrences per billion words). It was
proposed by Marc Brysbaert and colleagues as a standardized way to
compare word frequency, replacing older measures that were harder to
interpret consistently. A word scoring 6 shows up around once per
thousand words; one scoring 3 shows up around once per million. The gaps
between ranks in the table above come directly from that scale.
Grammatical data (gender, plural, and full case declension per noun) was
sourced from WiktionaryDE via the german_nouns package and joined
to the wordfreq rankings, then translated by hand. Details and the exact
regeneration pipeline are documented in this project’s tools/ directory
and public/data/words.js.
Why gender and case aren’t separated out
It would be simpler to teach der/die/das first, then circle back for case endings later. This app deliberately doesn’t do that. From a noun’s first appearance, its gender, its case forms (Nominativ, Akkusativ, Dativ, Genitiv), and both definite and indefinite articles are all in play together. Splitting them into stages would mean re-learning each noun several times instead of once, properly, and case marking in German leans on gender constantly, so practicing them apart doesn’t reflect how you’ll actually need to recall them.
Spaced repetition does the rest
Frequency ordering decides what you see first; a Leitner-box spaced repetition system decides when you see it again. Every noun/case/article combination you practice moves through a set of boxes: get it right enough times and it drifts toward “well known,” so it resurfaces less often. Struggle with it, or let it go too long unseen, and it comes back around sooner. You end up spending your practice time where it’s actually needed, instead of re-drilling things you already have down.
That’s the whole method: rank by real-world frequency, drill everything about a word together instead of staged, and let spaced repetition manage the schedule. Start practicing →
Zipf’s law explains why frequency ordering works mathematically. The practical payoff of following it is closer to the Pareto principle, a small share of the vocabulary doing most of the work. That’s next, below.
The 80/20 rule, applied to German vocabulary
The Pareto principle, named after economist Vilfredo Pareto, is the observation that in a lot of systems, roughly 80% of outcomes come from roughly 20% of causes. Pareto first noticed it looking at land ownership in Italy in 1896: about 80% of the land belonged to about 20% of the population. The ratio isn’t a law of nature and it isn’t always exactly 80/20, but the shape (a small input driving most of the output) turns up again and again, including in how language is actually used.
What this looks like in German, with real numbers
A Frequency Dictionary of German, by Randall Jones and Erwin Tschirner (Routledge, 2nd edition), ranks the 5,000 most common words in a 20-million-word corpus of German and measures how much of that corpus each rank actually covers:
- The 10 most frequent words alone account for roughly 27% of all running text.
- The next 10 bring that to roughly 35%.
- The full 5,000-word list covers between 84% and 95% of an average text, depending on genre.
That’s a strikingly small vocabulary doing most of the work, which is the 80/20 pattern showing up directly in text coverage data, not just as a loose analogy.
The same shape holds up outside German specifically. Research on lexical coverage in second-language reading (surveyed by vocabulary researcher Paul Nation and others) finds that the most frequent 1,000-2,000 word families in a language typically cover around 80% of an average written or spoken text. Reaching higher thresholds gets expensive fast: pushing coverage from 80% to the 95-98% usually considered necessary for comfortable, unassisted comprehension takes several thousand additional, much rarer words.
Zipf’s law and the Pareto principle aren’t the same thing
Zipf’s law is the specific mathematical pattern behind why this happens: word frequency in a language follows a rank-frequency curve where the Nth most common word shows up roughly 1/N as often as the most common one. The Pareto principle is the practical consequence for a learner: because that curve is so steep at the top, a small slice of vocabulary, ranked by frequency, covers a hugely disproportionate share of real usage.
How the curriculum applies it
Every noun unlocks in Zipf-rank order, so the curriculum is, by construction, always spending your time on the highest-coverage words available at your level. It’s the same reasoning A Frequency Dictionary of German is built around, applied to full case and article declension instead of just the bare word.
Frequency ordering isn’t a shortcut around learning German. It’s a reordering of the same 279 nouns (and growing) so the ones actually worth knowing well come first. Back to Zipf’s law above, or start practicing →.