The method

What is measured

The Project Gutenberg plain-text edition of each work, catalogued via gutendex. These are US public-domain works. Retrieved 2026-07-28.

No book text is reproduced anywhere on this site. Only computed statistics are published. Facts about a public-domain work carry no licence obligation, and Project Gutenberg's own licence applies to redistributing its texts — which this site does not do.

What is stripped before counting

Project Gutenberg wraps every text in a standard header and a long licence footer. Both are removed at the *** START OF *** and *** END OF *** markers before anything is counted, so Gutenberg's boilerplate never inflates an author's word count.

What remains is counted as published in that edition — including front matter such as prefaces, dedications, tables of contents and, in some editions, translator's notes. This is the single biggest reason a count here can differ from another source.

Why your number may differ — and both be right

Word counts for the same book vary by 5–10% between sources, routinely. The reasons are all legitimate:

• Different editions. Abridgements, translations and annotated editions genuinely contain different amounts of text.
• Front and back matter. Counting or excluding prefaces, appendices and notes moves the total substantially.
• What counts as a word. Hyphenated compounds, contractions and numerals are treated differently by different tools. Here a word is a run of letters, apostrophes and hyphens.

So a figure here is best read as "this specific edition, counted this specific way" — which is exactly why the method is on its own page rather than buried.

The readability formulas

Four standard formulas are computed: Flesch–Kincaid grade, Flesch reading ease, Gunning fog and Coleman–Liau. All four work from sentence length and word length; syllables are counted by a standard vowel-group heuristic, which is reliable in bulk and imperfect on individual unusual words.

Not Lexile. Lexile is a proprietary, trademarked measure owned by MetaMetrics. No Lexile score is published here, and the grade levels on this site are not Lexile equivalents.

All readability formulas share a deep limitation: they measure surface features only. They cannot detect irony, dialect, ambiguity or difficulty of ideas. A short-sentence modernist novel can score "easy" and be far harder to read than a Victorian one with long sentences.

Chapter counts

Chapters are detected from heading-shaped lines only, and are reported only when at least three are found. Where a book's structure cannot be read reliably from plain text, no chapter count is shown rather than a wrong one.

Reading time

Total words divided by a reading speed. The headline figure uses 250 words per minute, a common average for adult prose reading; every page also shows the full range from 150 to 450 wpm, because individual reading speed varies far more than single-figure estimates suggest.

Word counts and reading levels are computed from a specific public-domain edition and will differ from other editions and from other counting methods — see the method page.

Corrections

Spotted something wrong? Send the page URL.

SOURCE·Project Gutenberg full texts·retrieved 2026-07-28·REVIEWED JUL 2026