Portuguese text level checker

The level the checker gives is the level at which you would know 95% of the words in the text. That is the line reading research draws for reading without a dictionary: at 95%, about one word in twenty is new, which is one or two in a typical sentence, few enough to guess from the rest and still enjoy the text. The result says comfortable at that level.

It also gives two neighbours. Readable with effort is the level at which you would know 90%: one word in ten is new, so you can get through it with a dictionary and some patience, which is how most learners read their first real book. Easy is the level at which you would know 98%: almost nothing stops you, the kind of reading that builds speed and fluency (Nation, 2006).

The hardest word does not decide the level. Every real text has a few rare words: a name of a plant, a slang word, a term from someone's job. A single C2 word in a page of A2 Portuguese is a word to look up, not a reason to put the text away, and judging a text by its rarest word would put every newspaper beyond every learner. That is why the level is computed from the share of words at each level, and why the bar marks where 95% falls.

Rare words are those that are not among the 13,700 most common: shown in the C2 colour with a dotted line, and counted as above every level. Many are typos or words from another language, so the list of rare words is also a quick way to spot them. Names, of people and places, are shown in grey and left out of the level altogether, since knowing who Cida is has nothing to do with your Portuguese. So are laughs (kkkk, rsrs) and sounds like ah and hum.

Every word is placed by how common it is in Brazilian Portuguese, measured on the subtitles of Brazilian films and television: millions of lines of the Portuguese people actually speak, rather than the Portuguese of old novels. All the forms of a word count as one word, so falou, falamos and falando are all falar, and a contraction like do or pela is as easy as the two short words it joins.

The most common words do most of the work. The 1,000 most common make up about 80% of everything said in Brazilian film and television, and the 5,000 most common nearly 89% (the most common Portuguese words). After that each thousand words adds less, which is why the levels below get wider as they go up.

This word frequency checker uses the same levels as the free level test: a B1 learner in the test knows about 1,500 to 3,000 words, and a B1 word here is one ranked between 1,501 and 3,000. So a text the checker calls B1 is one a B1 reader can read comfortably.

The levels are a guide, in line with research on how many words learners know at each CEFR level (Milton & Alexiou, 2009), and the ranks run higher than those studies' counts because they count words as Brazilians say them.

Where frequency is a good guide, and where it isn't

Frequency is the best single predictor of whether a learner knows a word: the words you meet most are the ones you learn first, in any course and in any book. It is still a count, not a judgement of difficulty. Cognates with English, like informação, universidade and hospital, are easy for an English speaker to read whatever their rank. Topic words are common in their topic and rare everywhere else: goleiro (goalkeeper) is everywhere in a football report and nowhere in a novel. And a subtitle list knows conversation better than it knows writing, so formal words sit a little higher than a reader of the news would expect.

So treat the colours as a guide to what a text expects of a reader, not as a test of you. A word you already know is a word you already know, whatever its colour.

Anything in Brazilian Portuguese: a page of the book you want to read, the lyrics of a song you like, a paragraph of news, a chat with a Brazilian friend, the description of a video. The more text, the steadier the result; a page or two of a book gives a better answer than its first sentence, and a short message gives a rough one.

Each kind of text tells you something different. A page of a book tells you whether to start it now, start it with a dictionary, or come back to it in six months. Lyrics tell you how much of a song you can follow, and they often come out harder than you expect, because songs love rare and old-fashioned words. A news paragraph shows how much formal vocabulary you still need. A chat comes out easy, because texting is mostly the most common words written short.

Two examples, from the samples above:

Learn the words just above your level first, not the rarest ones. A word ranked 2,000 will turn up in the next text you read and the one after that; a word ranked 12,000 may not turn up again for months. That is why the list of words to learn is sorted by how common each word is, with how often it appears in your text beside it.

Set your level and the list shows only the words above it. Copy it, download it as a spreadsheet for Anki or Quizlet, or learn the whole band at once with the free Anki deck of the most common words, which starts at whichever level you choose. The 5,000 most common words are also on one page, each with its meaning and an example.

The easiest way to grow a vocabulary is to read a lot of text you can already mostly read. Graded readers are written for exactly that: our own, like Quem Abre a Porta, come in A2, B1 and B2 versions of the same story, so you can start with the one this checker calls comfortable and move up. The free books are sorted by level, from short graded stories to Machado de Assis.

If you have your own text, open it in the reader from the result: tap any word for its meaning in that sentence, hear it read aloud, and save the words you want to keep. And if you want a week of it, the 7-day reading challenge gives you a story a day.

All the free tools →