"Lorem ipsum dolor sit amet, consectetur adipiscing elit" is probably the most printed sentence nobody can read. Designers use it to fill a layout before the real words exist. This article covers where it actually comes from (less mysterious than the usual story), what placeholder text is supposed to do, and why filler in the language your page will really use catches problems that Latin hides. The examples come from our Lorem ipsum generator, which can write filler in Latin, Vietnamese, Spanish, Portuguese and Indonesian.
Where Lorem ipsum comes from
The text is a damaged quotation. Its source is Cicero's philosophical work De finibus bonorum et malorum ("On the ends of good and evil"), written in 45 BC. Sections 1.10.32 and 1.10.33 contain the sentence "Neque porro quisquam est, qui dolorem ipsum quia dolor sit amet, consectetur, adipisci velit…", roughly "nor is there anyone who loves pain itself because it is pain". Somebody cut it mid-word, so "dolorem" lost its first syllables and became "lorem", dropped and rearranged words, and added others. The result looks like Latin and means nothing.
The source was identified by Richard McClintock, a Latin scholar at Hampden-Sydney College in Virginia, who looked up "consectetur", a rare word, in classical literature and found Cicero. He described this in a 1994 letter to the design magazine Before & After.
You will often read that Lorem ipsum has been the printer's standard dummy text "since the 1500s". There is no documented example to support that date. What is documented is that the text was spread widely by Letraset dry-transfer sheets in the 1960s, and later shipped with desktop publishing software, which is how it reached every designer's toolbox.
What placeholder text is for
Filler has one job: to stand in for real text so that you judge the shape of a page, not its words. A reviewer looking at a mock-up full of real sentences starts editing the copy; with Lorem ipsum they look at the column width, the hierarchy, the spacing. That is still useful.
It has two classic risks:
- It ships. Search engines regularly index live pages with "Lorem ipsum" in them. Before any release, search the project for "lorem", and for whatever seed or marker you used.
- It flatters the design. Latin filler has a regular rhythm of medium-length words and no diacritics. Real text in most languages does not.
Why the language of the filler matters
Here are the same two sentences' worth of filler, generated with the seed demo, in three languages:
- Vietnamese: "Mùa thu buổi tối, mùa đông tìm kiếm sâu. Học nặng đứng, nhẹ buổi chiều con đường mưa." (18 words, 85 characters)
- Spanish: "Día después camino pequeño todos como cerrar algunos después café, quien cerrar camino. Abrir familia semana mercado lento camino después." (20 words, 138 characters)
- Indonesian: "Tenang datang pintu bisa, indah kenangan meja kenangan kerja sangat musim. Pendek membaca keluarga ini rendah hati." (17 words, 115 characters)
The words are real and correctly spelled; only their order is random. Three things show up immediately that Latin would hide.
Diacritics need vertical room. Vietnamese stacks marks above and below letters: "ở", "ữ", "ặ", "Ước". A heading set with a tight line height (1.0 or 1.1) can clip the top of "Ư" with a tone mark on it, and a font without full Vietnamese coverage will fall back to another font for some letters, which looks like a rendering bug. Latin filler never exercises this.
Word length changes line breaks. In this tool's word lists the average word is 6.2 letters in Latin, 5.4 in Indonesian, 4.9 in Spanish, 4.6 in Portuguese and 3.4 in Vietnamese, where nearly every word is a single syllable. A Vietnamese line breaks into many short pieces and rarely needs hyphenation; an Indonesian label such as "meninggalkan" (to leave) or "mengikuti" (to follow) is long enough to overflow a narrow button that "Lorem" fits in comfortably.
Counting words is language dependent. Vietnamese writes multi-syllable words with spaces inside: "thành phố" (city) is one word made of two syllables. Word counters, including ours, count space-separated tokens, so "thành phố" counts as two. The generator's "words" mode follows the same rule, so 12 words of Vietnamese filler is exactly 12 tokens ("Cũng mùa thu buổi tối mùa đông tìm kiếm cái bàn các."), even though a linguist would count fewer words.
A workflow that avoids surprises
- Mock up with filler in the target language, at realistic lengths: a product name of 2 to 4 words, a description of 2 sentences, an article paragraph of 4 to 6 sentences.
- Keep the seed. The generator's output is repeatable: the same seed and settings give the same text in any browser. Put the seed in the design file or the test fixture so screenshots stay comparable.
- Test the extremes. Generate the longest realistic version too (a long paragraph, a 7-word list item) and check that nothing overflows at a 390-pixel-wide phone screen.
- Replace before release and search for leftovers.
Common mistakes
- Testing only with Latin for a Vietnamese, Spanish or Indonesian site, then discovering clipped diacritics and broken buttons when the content arrives.
- Using machine gibberish ("ajsd kweo") as filler. It has no realistic word lengths or letters, so it tests nothing; real words in random order do.
- Leaving placeholder text in alt attributes and meta descriptions, where nobody sees it on the page but search engines do.
- Counting Vietnamese words by syllables in one tool and by meaning in another, then wondering why limits disagree. Decide which count your limit uses.
- Believing the 1500s story and repeating it. The Cicero source is documented; the date is not.
To check counts and character lengths of real copy once it arrives, the text toolkit counts words, characters and reading time.