LIX calculator

What a LIX calculator measures

LIX, short for the Swedish läsbarhetsindex, combines two visible properties of prose: average sentence length and the share of long words. If W is the number of words, S the number of sentences, and L the number of words longer than six letters, the formula is W ÷ S + 100 × L ÷ W. The result rises when sentences grow longer, when seven-letter words become more common, or both.

This page shows each part rather than returning a mystery number. The receipt lists W, S, L, the two formula terms, the unrounded result, and the nearest whole-number score. It also marks every long word and every punctuation run considered during sentence counting. You can therefore inspect why the number changed after an edit instead of guessing which hidden rule a calculator applied.

Why tokenization is part of the answer

The arithmetic is simple; deciding what counts is not. This calculator treats a maximal Unicode letter-or-number sequence as a word and keeps internal apostrophes, colons, and dashes attached. A compound such as e-postverktyg remains one token. Only its letters count toward the long-word threshold, so punctuation and digits cannot make a short word long. Accented and decomposed Unicode letters remain part of their word. For a separate receipt of grapheme clusters, code points and encoded length, use the character counter.

A run of periods, question marks, or exclamation marks can close one sentence. Decimal points are protected. A final fragment without punctuation counts as a sentence when it contains words. Those decisions are deliberate and visible in the sentence receipt. Other calculators may split compounds, count each mark in ?!, ignore a final fragment, or round differently; identical formulas can consequently produce different displayed scores.

Swedish and English presets

The preset changes one thing: the disclosed list of common abbreviations whose periods do not end a sentence. Swedish mode recognizes examples such as t.ex., bl.a., and m.fl.. English mode recognizes examples such as e.g., Mr., and etc.. Unknown abbreviations are not guessed. If one affects your text, its periods remain visible in the receipt so the limitation can be seen rather than silently hidden.

The formula itself does not change between modes. The familiar interpretation bands come from Swedish LIX practice and are useful descriptive context for Swedish prose. They are not validated American school grades, and an English result should not be presented as one. English readability tools often add syllable-based formulas such as Flesch or Flesch–Kincaid; this route calculates LIX only and says so explicitly.

Use the score as a prompt, not a verdict

A high score does not prove that writing is bad. A legal definition, research abstract, or technical instruction may need precise long terms. A low score does not prove that the text is accurate, coherent, or suitable for its reader. LIX cannot see structure, prior knowledge, typography, argument quality, or whether a short sentence is understandable. Use the marked tokens and sentence boundaries to find passages worth reading again.

If copied text contains broken PDF line wrapping or invisible characters, inspect it with the text cleaner first—but only if those artifacts are not part of what you intend to measure. For vocabulary repetition rather than readability, use the word frequency counter. Calculation and local file reading happen in this tab. No text is uploaded, logged, or stored.