The most common 5-letter words in five languages
Five-letter words are a small, arbitrary slice of a language, and since Wordle they are a famous one. Here are the twenty commonest in each of five languages, measured by how often they are said in film and television dialogue, along with the letters they use most, a good starting word for each, and the problem every non-English word game has to solve first: what to do about accents.
The twenty commonest, in five languages
| # | Spanish | French | Italian | Portuguese (Brazil) | German |
|---|---|---|---|---|---|
| 1 | estoy | était | tutto | muito | nicht |
| 2 | ahora | faire | della | estou | haben |
| 3 | vamos | comme | fatto | vamos | einen |
| 4 | bueno | votre | detto | minha | schon |
| 5 | tengo | cette | anche | então | jetzt |
| 6 | estás | quand | stato | agora | meine |
| 7 | tiene | alors | tutti | fazer | alles |
| 8 | puedo | chose | molto | tenho | keine |
| 9 | hacer | merci | posso | certo | ihnen |
| 10 | todos | juste | prima | disse | etwas |
| 11 | sabes | aussi | siamo | quero | bitte |
| 12 | puede | avoir | hanno | dizer | immer |
| 13 | usted | allez | delle | posso | warum |
| 14 | quién | autre | altro | sobre | sagen |
| 15 | nunca | temps | forse | mesmo | danke |
| 16 | dónde | notre | certo | coisa | gehen |
| 17 | favor | avait | tempo | todos | sehen |
| 18 | mucho | aller | nella | estão | einem |
| 19 | señor | monde | credo | tempo | diese |
| 20 | mejor | étais | fuori | assim | einer |
These are counted as five letters, so an accented letter is one letter: estás and était have five. Names (there are a lot of five-letter names in film dialogue — David, Harry, Sarah) and English left in the subtitles were removed by hand; nothing else was.
The lists are mostly verbs and the small words that hold sentences together, because that is what conversation is made of. Spanish opens with estoy, "I am"; Brazilian Portuguese with muito, "very, a lot"; French with était, "was"; Italian with tutto, "everything"; German with nicht, "not". Few are nouns. If you want words to learn rather than words to guess, that is the point: the commonest words of a language are its grammar.
Accents: the first decision every word game has to make
English Wordle has twenty-six letters and no accents. These five languages do not, and they differ a great deal in how much it matters.
| Language | Commonest letters, in order | Best starting words | Accented, of the top 100 | Pairs an accent-blind game merges |
|---|---|---|---|---|
| Spanish | a e o r s i n l | sería, serán, paseo | 13 | estás / estas, quién / quien, dónde / donde, están / estan |
| French | e r a i s t o n | serai, taire, étais | 12 | était / etait, votre / vôtre, pense / pensé, voilà / voila |
| Italian | a o e r i t s n | opera, amore, altro | 0 | città / citta, andra / andrà, caffè / caffe, verrà / verra |
| Portuguese (Brazil) | a o e r i s t n | serão, terão, estão | 8 | ideia / idéia, estas / estás |
| German | e t n a r s i h | raten, ernst, alten | 7 | schon / schön, würde / wurde, hatte / hätte, waren / wären |
The fourth column counts how many of each language's 100 commonest five-letter words carry an accent or another mark. The last column shows what happens if a game ignores them:
- Spanish and French lose real distinctions. Estás ("you are") and estas ("these") become one word, as do dónde and donde ("where?" and "where"). Among Spanish's 500 commonest five-letter forms, 17 pairs collapse.
- German loses the most meaning. Its umlauts mark different words, not just a different stress: schon is "already" and schön "beautiful", wurde is "was" and würde "would", hatte "had" and hätte "would have". 13 pairs among the 500 commonest merge.
- Italian barely notices. None of its 100 commonest five-letter words has an accent, because Italian writes one almost only on a final stressed vowel (città, caffè).
- Portuguese sits in between, and shows a spelling reform: ideia and idéia are the same word, spelt after and before the 1990 agreement that Brazil adopted in 2009.
Some of the merged pairs are also typing errors in the subtitles themselves — etait for était, andra for andrà — which says something about how often accents get dropped in practice.
The best starting words
A good Wordle opener uses five different letters that turn up often in the words you are trying to guess. So for each language we took the 500 commonest five-letter words, counted how many of them contain each letter (with accents stripped, but Spanish ñ and German ß kept as letters of their own), and scored every word with five different letters by how many of those words its letters reach. Candidates had to be in our own course word list, so that names and oddities could not win.
The winners: sería in Spanish, serai in French, opera in Italian, serão in Portuguese and raten in German. They lean on the same few letters, above all e, a and r: the Romance languages end most words in a vowel, and German a great many in -en.
Two cautions. This scoring is a simple rule of thumb, not a solver: a real solver also uses where each letter sits in the word, and the answer list of the game you are playing, which will not be our list. And a game may well choose its answers from a dictionary rather than from what people say, so the commonest words in speech are a good guide to likely answers, not a guarantee.
What the lists are good for
Treat the table as a vocabulary list in disguise: in every language, the commonest five-letter words are mostly verb forms and function words, and a learner who knows all twenty in a column has covered some of the most frequent words of that language. The full lists, in frequency order and with translations, are at Spanish, French, Italian, Portuguese and German, and the German vocabulary test shows in two minutes how far down its list you already are. The longest German words, rather than the shortest useful ones, are in the longest German words people actually use, and how fast each language's commonest words cover what is said is in which language has the steepest frequency curve.
Sources
- Lison, P. & Tiedemann, J. (2016). OpenSubtitles2016: Extracting Large Parallel Corpora from Movie and TV Subtitles. Proceedings of LREC 2016. aclanthology.org
- Dave, H. FrequencyWords: frequency lists from OpenSubtitles 2018, content licensed CC BY-SA 4.0. github.com
- Wikipedia: Portuguese-Language Orthographic Agreement of 1990 — the acute accent dropped from éi in words stressed on the second-last syllable (so idéia became ideia); Brazil changed on 1 January 2009.
- The names removed, the accent rule and the scoring are in
scripts/frequency_extras.py(FIVE_NOT_WORDS,five_letter).