Unique word counter and vocabulary range

A distinct spelling, after capitals are folded together. The and the are one word; cat and cats are two, and so are run and running, because collapsing them would mean guessing at word families and guessing wrongly in exactly the cases that matter. Apostrophes and internal hyphens do not split a word, so don't and state-of-the-art are one entry each, which keeps this total consistent with the unique-words figure the word counter reports for the same text.

The question has no answer without a length attached, which is the single most important thing to know about the measure. A 100-word paragraph will typically score somewhere around 70 per cent because almost every word in it is still new; the same author writing 5,000 words will score nearer 30, having said the and of a few hundred times. Nothing about the writing changed. That is why the length-corrected figure is on this page at all, and why any target ratio quoted without a word count beside it is meaningless.

Words that appear exactly once in a text — the term is Greek for “said once” and comes from classical scholarship, where a word occurring a single time in the surviving corpus is genuinely hard to translate. In ordinary writing they are unremarkable and abundant: in most texts of a few thousand words, something close to half the distinct vocabulary appears only once, and the proportion is stubbornly stable as the text grows. A low share is more interesting than a high one, because it means the piece keeps returning to the same words.

No. A high ratio can mean a rich vocabulary or it can mean an author reaching for a thesaurus and calling the same thing four different names, which is worse than repetition because the reader has to work out that all four are the same thing. Technical and legal writing deliberately repeats its key terms and scores low as a result; that is precision, not poverty. Read the ratio as a description of the text's texture, and then decide whether that texture suits what you are writing.

Because the grammar of English is finite and the subject matter is not. Every sentence you add needs the, a, of, is and to again, so the denominator keeps growing at full speed while the numerator slows down. This is not a flaw in your writing or in the measure — it is arithmetic, and it is exactly why the sliding-window figure exists. Watch that one instead: it stays roughly flat as a piece grows, and a real drop in it means the writing genuinely has become more repetitive.

Far fewer than they know. Estimates from large vocabulary studies put an average twenty-year-old American's known vocabulary at around 42,000 lemmas, and yet everyday speech and writing draw on a small fraction of that: a few thousand distinct words will carry an entire novel. Shakespeare's complete works run to about 884,000 words and use roughly 31,500 different ones, of which some 14,000 appear only once — which is a useful corrective to the idea that a big number in the box above is the goal.

Do it when you are comparing subject matter, and leave them in when you are measuring style. With the, of and their 125 relatives included you are measuring the whole texture of the prose, function words and all, which is what stylometry and authorship studies actually work from. With them set aside you are measuring the range of content vocabulary — closer to what a teacher means by vocabulary range in a piece of student writing. The two give noticeably different ratios for the same text, so pick one and stay with it across everything you compare.