Interlinear gloss formatter and checker (Leipzig Glossing Rules): aligned glosses, small caps, LaTeX gb4e and expex

What it does

Linguistics papers, theses and fieldwork reports show examples from other languages as interlinear glossed text: the original sentence, the same sentence split into morphemes, a gloss for every morpheme, and a free translation. The Leipzig Glossing Rules are the conventions most journals ask for. Getting the glosses to line up by hand, and keeping every hyphen in the gloss matched to a hyphen in the word, is fiddly. This tool checks your examples against the rules, shows them aligned, and writes them out for your document. Everything happens in your browser.

It works well with the publication plot maker, the corpus concordancer and the syntax tree drawer.

How to use it

1. Type one example per block, with a blank line between examples. The lines are, in order:
   • an optional title starting with # (the language, or a source),
   • the original line as written (untick “Each example starts with the original line” if you do not use one),
   • the morpheme line, with morphemes separated,
   • the gloss line, one gloss per word,
   • the free translation (quotes are optional; it may be left out).
2. Separate morphemes as the rules describe: - between affixes and stems, = before or after a clitic, ~ for reduplication, and < > around an infix (in both lines). When one morpheme needs several words in the gloss, join them with dots: give.IMP, go.out. A colon : or backslash \ marks a change that is not a separate morpheme and does not count as a boundary.
3. Write grammatical category labels in capitals (PL, 1SG, PST). They are shown in small capitals, as the rules suggest, and listed with their meanings. Labels not in the short built-in list are marked: add your own under Your own abbreviations as AV = actor voice.
4. Pick an export format and press Copy or Download file.

What is checked

• Word by word (rule 1): the morpheme line and the gloss line must have the same number of words.
• Morpheme by morpheme (rule 2): within each word, the gloss must have the same number of pieces, and the same kind of boundary in the same place (a hyphen against a hyphen, = against =, ~ against ~).
• Infixes (rule 9): the number of < > pieces must match.
• The original and morpheme lines usually have the same number of words; if not, you get a note, not an error, as splitting or joining words can be on purpose.

Words with a problem are outlined in red in the preview; hover or tap to see why.

Worked example

The two made-up examples in the box pass every check: 2 examples, 4 words, no problems. The Turkish one shows as Çocuk-lar okul-a git-ti over “child-pl school-dat go-pst” with pl, dat and pst in small caps, and the abbreviation list reads PL plural, DAT dative, PST past, IMP imperative, 1 first person, SG singular, 3 third person, M masculine, ACC accusative. Change go-PST to go and the check reports: Word 3: “git-ti.” has 2 morphemes but its gloss “go” has 1. The gb4e export for the first example is

\glll Çocuklar okula gitti.\\ then Çocuk-lar okul-a git-ti.\\, child-\textsc{pl} school-\textsc{dat} go-\textsc{pst}\\ and \trans `The children went to school.'.

FAQ

Does it gloss the language for me? No. It checks and formats glosses you write; it has no dictionary.

Which LaTeX package should I use? Both are widely used. gb4e is simple and common in journals; expex gives finer control over spacing and numbering. Use the one your template uses.

Why did my word turn small caps? Any gloss piece written fully in capitals (or a person number such as 1, 2, 3) is treated as a category label. Write lexical glosses in lower case.

Last updated 2026-11-09.