What Word Frequency Reveals About Your Writing

Paste a few paragraphs into a word frequency counter and something slightly embarrassing happens: your most-used words appear in a ranked list, and near the top sit the, of, and, a. That is normal — English runs on small connector words — but the list underneath the top tells you real things about your writing, your topic, and sometimes about language itself.

How a frequency counter counts

The rules matter more than they look, because different splitting rules produce different lists. Our Word Frequency Counter does this: it splits text on anything that is not a letter, digit, or apostrophe (so don't and 'tis stay whole words), lowercases everything by default (so The and the merge), ranks by count with ties broken alphabetically, and shows each word's share of the total.

Run the most famous line in English through it:

To be, or not to be, that is the question:
Whether 'tis nobler in the mind to suffer
The slings and arrows of outrageous fortune.

25 words, 20 unique
to    3   (12%)
the   3   (12%)
be    2   (8%)
'tis, and, arrows, fortune...  1 each

Note 'tis counting as one word — the apostrophe rule earning its keep in a 400-year-old text. The percentage is simply count divided by total words.

Unique words and the variety number

The counter also reports 20 unique words out of 25 total — a ratio of 0.8. That ratio (linguists call it the type–token ratio) measures vocabulary variety, but it comes with a catch: it sinks as text gets longer, because you cannot keep introducing new words while the keeps showing up. A 25-word sample at 0.8 and a 25,000-word report at 0.15 are not comparable. Compare variety only between texts of similar length, and treat a low ratio in a short piece as a signal worth investigating — often it means one idea is being restated with the same vocabulary.

Zipf's law and the function-word wall

Word frequencies follow a strange regularity called Zipf's law: a word's frequency is roughly proportional to 1 divided by its rank. The number-one word appears about twice as often as number two, which appears about twice as often as number three, and so on. Real corpora approximate this uncannily well.

The top of the ranking is always the same wall of function words — the, of, and, to, a, in, that — grammatical glue that carries almost no topic meaning. In large English corpora, the alone is roughly 6–7% of all words. So the useful editing move is to skim past the wall: the words ranked between about 20 and 200 are your actual subject matter, and the long tail of once-only words is where typos and strained vocabulary live. Frequency lists are also how keyword stuffing gets caught — a page about coffee where coffee climbs to a quarter of all words reads like spam to readers and search engines alike.

Letters keep their own league table

Count the 26 single-character words instead and you get letter frequency, the older sibling of word frequency. In running English text, the letter e leads at roughly 12.7% of all letters, followed by t (about 9%) and a (about 8%) — figures that come from classic cryptanalysis tables and have stayed stable for a century. The Letter Frequency Counter computes this for any text you paste.

A favorite demo: the pangram "the quick brown fox jumps over the lazy dog" contains all 26 letters, yet its letters still refuse to be fair — o wins with 4 of 35 letters (11.43%) and e is second with 3 (8.57%). Being a pangram does not flatten the distribution. Letter frequency powers crosswords and word games, and historically it broke ciphers: in a simple substitution cipher, the most common ciphertext letter is usually e in disguise. It is even etched into old technology — Morse code gives e the shortest symbol of all, a single dot, precisely because it is the letter a message needs most often.

From counts to edits

Totals first, texture second. The Word Counter gives you the frame — words, sentences, paragraphs, reading time at 200 words per minute — while frequency gives the texture inside it. Three checks worth running on your own drafts:

• Read your top ten. Anything there more often than you intended? Unintentional pet words are invisible while writing and glaring in a frequency list.
• Hunt echo words. The same word twice in one sentence reads as a mistake even when each use is fine on its own.
• Watch the wall. If a flabby paragraph's visible words are mostly the, of, and was, the sentence structure — not the vocabulary — is the problem.

The percentages are ordinary arithmetic — a count divided by a total, expressed as a share, the same everyday math behind averages and proportions (see Mean vs Median vs Mode: How Averages Work).

Try it yourself

Paste a few paragraphs of your own writing into the Word Frequency Counter and read the top ten as a diagnosis, not a ranking. Then run your text through the Letter Frequency Counter to see how closely your letters track English's e-t-a order. And keep the Word Counter open for totals and reading time while you edit — the frame around the texture.