How many Egyptian Arabic words do you actually need?

The usual figure quoted for a language is that 1,000 words gets you 80% of ordinary speech. I wanted to check it against Egyptian Arabic specifically, so I ran the numbers on the frequency list this site is built from: 2,404 words drawn from a corpus of 29.3 million tokens of written Egyptian.

The headline number holds up almost exactly. Everything interesting is underneath it.

The curve

Ten words cover a sixth of everything. A hundred cover close to half. The thousand-word figure lands at 83%, and the last 1,400 words on the list buy you the remaining 17%.

One caveat that matters, because most articles quoting numbers like these skip it: these percentages are shares of the tokens this list accounts for, not of all Egyptian Arabic ever written. The list stops at 2,404 words, so the real corpus is bigger and the true coverage of running text at any rank is somewhat lower than the figure above. The shape of the curve is right. Treat the exact numbers as an upper bound.

Now look at what the first twenty words actually are

There is nothing there you can study. 52 of the top 100 are function words — prepositions, pronouns, particles, conjunctions. They carry grammar, not meaning. You cannot usefully put اللي on a flashcard, because knowing that it means "who" or "which" tells you nothing about how to use it; you learn it by seeing it four hundred times in sentences until the shape of a relative clause feels normal.

The share drops as you go down the list — 52% in the first hundred, 37% between 100 and 300, 23% between 300 and 500 — and that decline is the actual answer to "where do I start studying vocabulary". Not at the top.

The corpus has a personality

Four of the top 50 words are invocations of God: الله, والله, يارب, ربنا. حبيبي ("my dear") sits at rank 100. قلبي ("my heart") is inside the top 30.

That is not what a transcript of people buying vegetables looks like. It is what song lyrics and social media look like, which is where this kind of corpus generally comes from. The religious vocabulary is genuinely high-frequency in real Egyptian speech — إن شاء الله and الحمد لله punctuate ordinary conversation constantly — but the romantic vocabulary is inflated by the source material.

There is straightforward noise in there too. Rank 15 is علي, glossed as the name Ali, which is almost certainly على ("on") written without its final hamza and counted separately. Any frequency list assembled from raw text has artefacts like this. It is worth knowing they exist before you treat rank order as gospel.

So what should you actually do

Do not study the first hundred words as vocabulary. Read and listen enough that they become invisible. They will, quickly, because they are in every sentence.

Start deliberate study somewhere around rank 150. That is roughly where content words start outnumbering grammar words, and where a flashcard begins to be worth making.

Switch from frequency to topics after the first few hundred. Past that point, frequency rank stops predicting usefulness. Whether مطار ("airport") is word 800 or word 1,600 matters much less than whether you are about to fly to Cairo. This is why the vocabulary on this site is grouped by subject rather than served as one long ranked list.

Expect the curve to flatten and keep going anyway. Getting from 83% to 90% takes another 500 words, and that last stretch is where the difference between following a conversation and participating in one actually lives.

Try it

The top of the frequency list is on the site, ranked and with the raw occurrence counts, so you can see the shape of it for yourself rather than take my word for it.

• The 500 most common Egyptian Arabic words — ranked, with occurrence counts.
• Vocabulary by topic — the useful band, sorted by subject instead.
• Egyptian Arabic for beginners — where vocabulary fits in the wider order of things.